Department of Psychology (Invited): Cognitive Brownbag Series, Colorado State University, CO.
The popularity of machine learning, specifically neural networks, is rapidly growing in nearly all areas of science. The explosion of these methods also coincides with a growing influx of computationally expensive data sets and the need for high efficiency in solving predication problems. However, there is also some hesitancy in adopting the use of neural networks due to concerns about their reliability, reproducibility, and interpretability – thus, they are often described as “black boxes.”
In climate science, we often consider signal-to-noise problems to help disentangle human-caused climate change from natural variability. These applications typically involve complicated nonlinear relationships between different feedbacks at play in the ocean, land, and atmosphere. Recent work has shown that neural networks can be a promising tool for solving these types of statistical problems when combined with explainability techniques developed by the fields of computer science and image processing. Interestingly, these methods have revealed that neural networks often leverage regional patterns of climate change indicators in order to make their predictions. In this talk, I will share examples from climate science that use a few of these visualization methods to peer into the “black box” of neural networks, which help us to better understand their decision-making processes while also learning new science. The same machine learning visualization methods can be easily adapted for a wide variety of applications and other scientific fields of study.
Exploring climate change signals with explainable AIZachary Labe
This document provides an overview of using machine learning and explainable AI techniques to analyze climate model data and make predictions about climate signals and forcings. It discusses:
1) The growing use of machine learning tools like neural networks in weather and climate research to better understand models, find relationships in data, and make predictions.
2) An example of using a neural network to classify decade from surface temperature maps and explaining the predictions using Layer-wise Relevance Propagation.
3) Results showing the neural network had greater skill in identifying patterns related to greenhouse gas forcing versus aerosol forcing, and LRP helped identify the relevant regions driving these predictions.
Assessing climate variability and change with explainable neural networksZachary Labe
I. Explainable neural networks can reveal patterns of climate change in large ensembles simulated with different combinations of external forcing.
II. Neural networks can identify unique model differences and biases between large ensembles and observations.
III. The presentation raises the question of what neural networks can reveal about predictability in the climate system, using the example of the debated "global warming hiatus" period.
Climate Signals in CESM1 Single-Forcing Large Ensembles Revealed by Explainab...Zachary Labe
26th Annual CESM Workshop - Machine Learning: CESM-Related Efforts
In this study, we use an explainable artificial intelligence met hod to identify climate signals that are found in a new set of single-forcing large ensembles from CESM1. To compare patterns between simulations, we adopt an artificial neural network (ANN) that predicts the year from input maps of near-surface temperature. We find that the North Atlantic Ocean is an important region for the ANN to make its prediction, especially for the simulation forced without time-evolving industrial aerosols.
Using explainable neural networks for comparing climate model projectionsZachary Labe
1. The document discusses using explainable neural networks to compare climate model projections and evaluate which climate models best match observations.
2. Temperature maps from observations are input into a neural network trained on climate model data to classify each observation year with a climate model.
3. Layer-wise relevance propagation is used to explain the neural network's classifications and identify differences between climate models, which can help evaluate models, especially in regions with known biases like the Arctic.
Revealing climate change signals with explainable AIZachary Labe
1. The document discusses using explainable artificial intelligence methods to analyze climate model data and reveal patterns of climate change.
2. An artificial neural network was trained to predict decades from maps of surface temperature and showed greater skill at predicting decades when greenhouse gases were the only forcing compared to when aerosols were the only forcing.
3. Explainable AI techniques showed the North Atlantic region was particularly important for the neural network's predictions in climate model experiments forced by both aerosols and greenhouse gases.
Disentangling Climate Forcing in Multi-Model Large Ensembles Using Neural Net...Zachary Labe
The relative roles of individual forcings on large-scale climate variability remain difficult to disentangle within fully-coupled global climate model simulations. Here, we train an artificial neural network (ANN) to classify the climate forcings of a new set of CESM1 initial-condition large ensembles that are forced by different combinations of aerosol (industrial and biomass burning), greenhouse gas, and land-use/land-cover forcings. As a result of learning the regional responses of internal variability to the different external forcings, the ANN is able to successfully classify the dominant forcing for each model simulation. Using recently developed explainable AI methods, such as layerwise relevance propagation, we then compare the patterns of climate variability identified by the ANN between different external climate forcings that are learned by the neural network. Further, we apply this ANN architecture on additional climate simulations from the multi-model large ensemble archive, which include all anthropogenic and natural radiative forcings. From this collection of initial-condition ensembles, the ANN is also able to detect changes in atmospheric internal variability between the 20th and 21st centuries by training on climate fields after the mean forced signal has already been removed. This ANN framework and its associated visualization tools provide a novel approach to extract complex patterns of observable and projected climate variability and trends in Earth system models. (from https://ams.confex.com/ams/101ANNUAL/meetingapp.cgi/Paper/379553)
Exploring climate change signals with explainable AIZachary Labe
This document provides an overview of using machine learning and explainable AI techniques to analyze climate model data and make predictions about climate signals and forcings. It discusses:
1) The growing use of machine learning tools like neural networks in weather and climate research to better understand models, find relationships in data, and make predictions.
2) An example of using a neural network to classify decade from surface temperature maps and explaining the predictions using Layer-wise Relevance Propagation.
3) Results showing the neural network had greater skill in identifying patterns related to greenhouse gas forcing versus aerosol forcing, and LRP helped identify the relevant regions driving these predictions.
Assessing climate variability and change with explainable neural networksZachary Labe
I. Explainable neural networks can reveal patterns of climate change in large ensembles simulated with different combinations of external forcing.
II. Neural networks can identify unique model differences and biases between large ensembles and observations.
III. The presentation raises the question of what neural networks can reveal about predictability in the climate system, using the example of the debated "global warming hiatus" period.
Climate Signals in CESM1 Single-Forcing Large Ensembles Revealed by Explainab...Zachary Labe
26th Annual CESM Workshop - Machine Learning: CESM-Related Efforts
In this study, we use an explainable artificial intelligence met hod to identify climate signals that are found in a new set of single-forcing large ensembles from CESM1. To compare patterns between simulations, we adopt an artificial neural network (ANN) that predicts the year from input maps of near-surface temperature. We find that the North Atlantic Ocean is an important region for the ANN to make its prediction, especially for the simulation forced without time-evolving industrial aerosols.
Using explainable neural networks for comparing climate model projectionsZachary Labe
1. The document discusses using explainable neural networks to compare climate model projections and evaluate which climate models best match observations.
2. Temperature maps from observations are input into a neural network trained on climate model data to classify each observation year with a climate model.
3. Layer-wise relevance propagation is used to explain the neural network's classifications and identify differences between climate models, which can help evaluate models, especially in regions with known biases like the Arctic.
Revealing climate change signals with explainable AIZachary Labe
1. The document discusses using explainable artificial intelligence methods to analyze climate model data and reveal patterns of climate change.
2. An artificial neural network was trained to predict decades from maps of surface temperature and showed greater skill at predicting decades when greenhouse gases were the only forcing compared to when aerosols were the only forcing.
3. Explainable AI techniques showed the North Atlantic region was particularly important for the neural network's predictions in climate model experiments forced by both aerosols and greenhouse gases.
Disentangling Climate Forcing in Multi-Model Large Ensembles Using Neural Net...Zachary Labe
The relative roles of individual forcings on large-scale climate variability remain difficult to disentangle within fully-coupled global climate model simulations. Here, we train an artificial neural network (ANN) to classify the climate forcings of a new set of CESM1 initial-condition large ensembles that are forced by different combinations of aerosol (industrial and biomass burning), greenhouse gas, and land-use/land-cover forcings. As a result of learning the regional responses of internal variability to the different external forcings, the ANN is able to successfully classify the dominant forcing for each model simulation. Using recently developed explainable AI methods, such as layerwise relevance propagation, we then compare the patterns of climate variability identified by the ANN between different external climate forcings that are learned by the neural network. Further, we apply this ANN architecture on additional climate simulations from the multi-model large ensemble archive, which include all anthropogenic and natural radiative forcings. From this collection of initial-condition ensembles, the ANN is also able to detect changes in atmospheric internal variability between the 20th and 21st centuries by training on climate fields after the mean forced signal has already been removed. This ANN framework and its associated visualization tools provide a novel approach to extract complex patterns of observable and projected climate variability and trends in Earth system models. (from https://ams.confex.com/ams/101ANNUAL/meetingapp.cgi/Paper/379553)
The Enge Focal Plane detector needs to be recommissioned to improve measurements of important nuclear reactions for astrophysics. The detector previously provided improved energy resolution and angular information but required repairs. The team worked on software to analyze signals from the detector and calibrated it using a 60Co source, though ultimately no data could be collected. Recommissioning would allow remeasuring reactions relevant to nova nucleosynthesis and abundance patterns in stars.
This document summarizes an experiment that measured the interior air temperature of a model doghouse facing south and west at different times of day.
In the morning (simulating winter conditions), the south-facing doghouse had a higher interior air temperature, while the west-facing doghouse was cooler. In the late afternoon (simulating summer), the south-facing doghouse was cooler, while the west-facing interior was warmer.
The results demonstrate that a structure oriented with its long side running east-west performs better for passive solar heating in winter and cooling in summer compared to other orientations. Some limitations included variable weather conditions and simplifications in the thermal model.
Burlap, cheesecloth, and window screen were tested as canopy materials to reduce soil heating from solar radiation. Irradiance measurements found burlap allowed the least sunlight, followed by window screen then cheesecloth. Soil surface and depth temperatures followed this trend, with burlap maintaining the coolest temperatures. COMSOL modeling of heat flux agreed with experimental data. In conclusion, burlap most effectively reduced soil temperatures at all depths tested, demonstrating its utility for greenhouse shade coverings.
1. The document outlines the class schedule for an advanced soil and water engineering course, covering topics like physical characteristics of soil, micrometeorology, soil water, and applications of soil and water engineering.
2. It discusses the concepts of surface radiation balance and heat transfer, including equations for net radiation, solar radiation, reflected radiation, and outgoing longwave radiation.
3. It explains the components of surface heat balance - net radiation, ground heat flux, sensible heat flux, and latent heat flux - and the equations governing heat and vapor transfer through atmospheric resistances.
The document discusses the use of full-range spectroradiometers for measuring solar spectral energy. ASD spectroradiometers are portable instruments that measure wavelengths from 350-2500 nm, allowing for calculation of parameters like aerosol optical depth. They are optimally designed for faster field measurements and calibrated for radiometric collection. ASD instruments offer a solution for analyzing and measuring solar spectral energy with accessories like diffuse fore optics for measuring full sky irradiance.
Substantial disagreement continues between modeling studies in attributing midlatitude climate extremes to Arctic sea-ice anomalies. This is a result of uncertainties due to internal variability, nonlinear interactions, model biases, or more likely a combination of these effects. In this study, we use large ensembles from two high-top atmospheric general circulation models (SC-WACCM4 and E3SM) to separate the sea ice-forced signal from atmospheric internal variability (noise). Following protocol for the Polar Amplification Model Intercomparison Project (PAMIP), each simulation is prescribed with either pre-industrial, present-day, or future levels of sea-ice concentration, which are associated with global warming projections of 2°C. We use 300 ensemble members per simulation to obtain large sample sizes for robust statistics in the context of internal variability.
While an equatorward shift of the eddy-driven jet is found in boreal winter, the response to future sea-ice loss is small relative to climatology and highly sensitive to the number of ensemble members considered. On average, a sea ice-forced signal in the large-scale circulation cannot be distinguished from atmospheric internal variability in our simulations. A low signal-to-noise ratio is also demonstrated in the stratosphere, where the sign of the polar vortex response can be interpreted differently depending on the ensemble size. However, the local thermodynamic effects are statistically significant with strong surface warming and increases in precipitation found in the vicinity of newly ice-free areas. This warming is generally confined to the Arctic, and there is little response in the midlatitudes. Our results highlight the important role of internal variability in the extratropics and emphasize the need for especially large ensembles (>150-200 members) when assessing the dynamical response to both present-day and future Arctic sea-ice loss. (from https://ams.confex.com/ams/2020Annual/meetingapp.cgi/Paper/367289)
This study compared the convection coefficients of two potato slices cooling via free and forced convection. One slice cooled freely on a rack while the other was cooled with a standing fan blowing air over it. Temperature probes measured the internal temperature of each slice over time. The slice under forced convection cooled faster, with its convection coefficient calculated as 20 W/(m2K) compared to 9.5 W/(m2K) for the freely cooling slice. A COMSOL model also showed the internal temperature distribution over time matched the measured data. The experiment validated that forced convection enhances heat transfer compared to free convection.
1) UAVSAR collected repeat-pass SAR data over agricultural fields in Canada during a soil moisture campaign to analyze the effect of soil moisture on polarimetric and interferometric measurements.
2) Both backscatter and interferometric phase correlations decreased as soil moisture changed more between passes, indicating soil moisture impacts the SAR measurements.
3) Models captured general backscatter and polarimetric phase trends with soil moisture but underestimated the observed variation in interferometric phase with changes in moisture.
This document describes how rainfall estimates are created from raw radar data at the National Centers for Environmental Information (NCEI). It explains that multiple radars are combined and merged to get nationwide coverage with 2 minute resolution on a standardized grid. Raw radar returns are cleaned of contamination from birds and other sources. The vertical radar scans are integrated into a single rainfall estimate. The final product provides a consistent, high quality rainfall estimate compared to the original radar data which had poorer spatial and temporal resolution and quality issues.
This document summarizes a presentation analyzing hourly temperature and energy data from Barrow, Alaska between 1985-2017. A model was created that found: 1) CO2 concentrations played an important role in explaining the positive net inward energy imbalance. 2) Temperature data was best explained by including CO2 concentrations as an explanatory variable, rather than alternative drivers. 3) The model accurately predicted out-of-sample temperature and energy data, supporting the role of CO2 in increasing temperatures. The analysis provides evidence that contradicts views of climate change deniers.
Climate statisticians analyze observational climate data and model simulations to detect trends, attribute causes, and quantify uncertainties. They use statistical methods like linear regression to attribute observed warming to human and natural factors. They also use extreme value theory to describe rare weather events and project how these extremes may change with continued warming. A key task is quantifying various sources of uncertainty in climate projections, like different model sensitivities and emissions scenarios.
The document discusses an ensemble generation scheme for ocean analysis systems that aims to provide uncertainty estimates of the ocean state. The scheme involves randomly perturbing both assimilated observations and surface forcing inputs to generate an ensemble of ocean state estimates. This allows the analysis system to account for uncertainties in observations and surface forcing when assessing the ocean state. The project developing this scheme is taking place from 2018 to 2020 at ECMWF.
Dynamic shading of skylights can improve indoor environments. A field experiment was conducted in an office space to test different shading devices on skylights. Skylights were separated to test performance individually. Sensors measured temperature, light, and glare levels with and without shading. Results showed that shading can reduce heat gain and glare while allowing daylight, and provide insulation at night. Computer simulations confirmed that roller shades and angled blinds effectively reduce glare from skylights while maintaining sufficient light levels. An optimal control strategy is being developed to automatically adjust shading based on weather and sunlight conditions.
This document discusses using spectral corrections to better model PV performance. It presents an overview of how the spectral distribution of sunlight changes over the day and affects PV module performance. Models that account for the spectral response of modules and changing spectral irradiance can provide more accurate estimates of performance metrics like current and power output. The document shares results of analyzing spectral data that demonstrate variations in average module response under direct normal, diffuse horizontal, and global irradiance over the course of a day due to these spectral effects. Future work aims to refine estimates of air mass effects and integrate spectral modeling into broader PV modeling.
The document examines the relationship between cloud droplet effective radius and cloud top height in deep convective clouds using CloudSat data. It finds that as cloud droplet effective radius decreases, cloud top height increases, supporting the hypothesis that aerosols can invigorate deep convective clouds. However, this negative correlation is not seen for clouds with lower cloud top heights. The relationship also depends on surface temperature, with the threshold altitude for a negative correlation increasing at higher temperatures.
In today's world, natural disasters are an ever increasing presence in our lives and working environments. SolSpec has developed a unique capacity for utilizing aerial imagery to prevent and mitigate disasters.
Machine learning for evaluating climate model projectionsZachary Labe
IEEE-Student Branch- IIT Indore, Tech-Talks 2.0. Remote Presentation (Dec 2022) (Invited).
The popularity of machine learning methods, such as neural networks, is rapidly expanding in nearly all areas of science. The interest in these tools also coincides with a growing influx of big data and the need for high efficiency in solving predication problems. However, there is also some hesitancy in adopting the use of neural networks due to concerns about their reliability, reproducibility, and interpretability.
In climate science, we often consider signal-to-noise problems to help disentangle human-caused climate change from natural variability. These applications typically involve complicated relationships between different feedbacks at play in the ocean, cryosphere, land, and atmosphere. Recent work has shown that neural networks can be a promising tool for solving these types of statistical problems when combined with explainability techniques developed by the fields of computer science and image processing. Interestingly, these methods have revealed that neural networks often leverage regional patterns of climate change in order to make their predictions. In this webinar, I will share examples from climate science that use a few of these visualization methods to peer into the “black box” of neural networks, which help us to better understand their decision-making process while also learning new science. The same machine learning visualization methods can be easily adapted for a wide variety of applications and other scientific fields of study.
Using explainable machine learning for evaluating patterns of climate changeZachary Labe
22 February 2023…
Natural Sciences Group Seminar (Presentation): Using explainable machine learning for evaluating patterns of climate change, Washington State University Vancouver, USA. Remote Presentation.
References:
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464
Labe, Z.M. and E.A. Barnes (2022), Predicting slowdowns in decadal climate warming trends with explainable neural networks. Geophysical Research Letters, DOI:10.1029/2022GL098173
Exploring explainable machine learning for detecting changes in climateZachary Labe
9 February 2023…
Department of Earth, Ocean, and Atmospheric Science Colloquium (Presentation): Exploring explainable machine learning for detecting changes in climate, Florida State University, USA. Remote Presentation.
References:
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464
Labe, Z.M. and E.A. Barnes (2022), Predicting slowdowns in decadal climate warming trends with explainable neural networks. Geophysical Research Letters, DOI:10.1029/2022GL098173
Po-Chedley, S., J.T. Fasullo, N. Siler, Z.M. Labe, E.A. Barnes, C.J.W. Bonfils, and B.D. Santer (2022). Internal variability and forcing influence model-satellite differences in the rate of tropical tropospheric warming. Proceedings of the National Academy of Sciences, DOI:10.1073/pnas.2209431119
Learning new climate science by thinking creatively with machine learningZachary Labe
Presentation for: GFDL/AOS Summer Internship Lecture Series
The popularity of machine learning is rapidly growing in nearly all areas of Earth science. However, there is also some hesitancy in adopting the use of these methods due to concerns about their reliability, reproducibility, and interpretability – thus, they are often described as “black boxes.”
In this talk, I will introduce a few simple examples from climate science that leverage new visualization methods to peer into the machine learning “black box,” which help us to better understand their predictions while also learning new science. These same machine learning visualization tools can be easily adapted for a wide variety of applications and other scientific fields of study.
The Enge Focal Plane detector needs to be recommissioned to improve measurements of important nuclear reactions for astrophysics. The detector previously provided improved energy resolution and angular information but required repairs. The team worked on software to analyze signals from the detector and calibrated it using a 60Co source, though ultimately no data could be collected. Recommissioning would allow remeasuring reactions relevant to nova nucleosynthesis and abundance patterns in stars.
This document summarizes an experiment that measured the interior air temperature of a model doghouse facing south and west at different times of day.
In the morning (simulating winter conditions), the south-facing doghouse had a higher interior air temperature, while the west-facing doghouse was cooler. In the late afternoon (simulating summer), the south-facing doghouse was cooler, while the west-facing interior was warmer.
The results demonstrate that a structure oriented with its long side running east-west performs better for passive solar heating in winter and cooling in summer compared to other orientations. Some limitations included variable weather conditions and simplifications in the thermal model.
Burlap, cheesecloth, and window screen were tested as canopy materials to reduce soil heating from solar radiation. Irradiance measurements found burlap allowed the least sunlight, followed by window screen then cheesecloth. Soil surface and depth temperatures followed this trend, with burlap maintaining the coolest temperatures. COMSOL modeling of heat flux agreed with experimental data. In conclusion, burlap most effectively reduced soil temperatures at all depths tested, demonstrating its utility for greenhouse shade coverings.
1. The document outlines the class schedule for an advanced soil and water engineering course, covering topics like physical characteristics of soil, micrometeorology, soil water, and applications of soil and water engineering.
2. It discusses the concepts of surface radiation balance and heat transfer, including equations for net radiation, solar radiation, reflected radiation, and outgoing longwave radiation.
3. It explains the components of surface heat balance - net radiation, ground heat flux, sensible heat flux, and latent heat flux - and the equations governing heat and vapor transfer through atmospheric resistances.
The document discusses the use of full-range spectroradiometers for measuring solar spectral energy. ASD spectroradiometers are portable instruments that measure wavelengths from 350-2500 nm, allowing for calculation of parameters like aerosol optical depth. They are optimally designed for faster field measurements and calibrated for radiometric collection. ASD instruments offer a solution for analyzing and measuring solar spectral energy with accessories like diffuse fore optics for measuring full sky irradiance.
Substantial disagreement continues between modeling studies in attributing midlatitude climate extremes to Arctic sea-ice anomalies. This is a result of uncertainties due to internal variability, nonlinear interactions, model biases, or more likely a combination of these effects. In this study, we use large ensembles from two high-top atmospheric general circulation models (SC-WACCM4 and E3SM) to separate the sea ice-forced signal from atmospheric internal variability (noise). Following protocol for the Polar Amplification Model Intercomparison Project (PAMIP), each simulation is prescribed with either pre-industrial, present-day, or future levels of sea-ice concentration, which are associated with global warming projections of 2°C. We use 300 ensemble members per simulation to obtain large sample sizes for robust statistics in the context of internal variability.
While an equatorward shift of the eddy-driven jet is found in boreal winter, the response to future sea-ice loss is small relative to climatology and highly sensitive to the number of ensemble members considered. On average, a sea ice-forced signal in the large-scale circulation cannot be distinguished from atmospheric internal variability in our simulations. A low signal-to-noise ratio is also demonstrated in the stratosphere, where the sign of the polar vortex response can be interpreted differently depending on the ensemble size. However, the local thermodynamic effects are statistically significant with strong surface warming and increases in precipitation found in the vicinity of newly ice-free areas. This warming is generally confined to the Arctic, and there is little response in the midlatitudes. Our results highlight the important role of internal variability in the extratropics and emphasize the need for especially large ensembles (>150-200 members) when assessing the dynamical response to both present-day and future Arctic sea-ice loss. (from https://ams.confex.com/ams/2020Annual/meetingapp.cgi/Paper/367289)
This study compared the convection coefficients of two potato slices cooling via free and forced convection. One slice cooled freely on a rack while the other was cooled with a standing fan blowing air over it. Temperature probes measured the internal temperature of each slice over time. The slice under forced convection cooled faster, with its convection coefficient calculated as 20 W/(m2K) compared to 9.5 W/(m2K) for the freely cooling slice. A COMSOL model also showed the internal temperature distribution over time matched the measured data. The experiment validated that forced convection enhances heat transfer compared to free convection.
1) UAVSAR collected repeat-pass SAR data over agricultural fields in Canada during a soil moisture campaign to analyze the effect of soil moisture on polarimetric and interferometric measurements.
2) Both backscatter and interferometric phase correlations decreased as soil moisture changed more between passes, indicating soil moisture impacts the SAR measurements.
3) Models captured general backscatter and polarimetric phase trends with soil moisture but underestimated the observed variation in interferometric phase with changes in moisture.
This document describes how rainfall estimates are created from raw radar data at the National Centers for Environmental Information (NCEI). It explains that multiple radars are combined and merged to get nationwide coverage with 2 minute resolution on a standardized grid. Raw radar returns are cleaned of contamination from birds and other sources. The vertical radar scans are integrated into a single rainfall estimate. The final product provides a consistent, high quality rainfall estimate compared to the original radar data which had poorer spatial and temporal resolution and quality issues.
This document summarizes a presentation analyzing hourly temperature and energy data from Barrow, Alaska between 1985-2017. A model was created that found: 1) CO2 concentrations played an important role in explaining the positive net inward energy imbalance. 2) Temperature data was best explained by including CO2 concentrations as an explanatory variable, rather than alternative drivers. 3) The model accurately predicted out-of-sample temperature and energy data, supporting the role of CO2 in increasing temperatures. The analysis provides evidence that contradicts views of climate change deniers.
Climate statisticians analyze observational climate data and model simulations to detect trends, attribute causes, and quantify uncertainties. They use statistical methods like linear regression to attribute observed warming to human and natural factors. They also use extreme value theory to describe rare weather events and project how these extremes may change with continued warming. A key task is quantifying various sources of uncertainty in climate projections, like different model sensitivities and emissions scenarios.
The document discusses an ensemble generation scheme for ocean analysis systems that aims to provide uncertainty estimates of the ocean state. The scheme involves randomly perturbing both assimilated observations and surface forcing inputs to generate an ensemble of ocean state estimates. This allows the analysis system to account for uncertainties in observations and surface forcing when assessing the ocean state. The project developing this scheme is taking place from 2018 to 2020 at ECMWF.
Dynamic shading of skylights can improve indoor environments. A field experiment was conducted in an office space to test different shading devices on skylights. Skylights were separated to test performance individually. Sensors measured temperature, light, and glare levels with and without shading. Results showed that shading can reduce heat gain and glare while allowing daylight, and provide insulation at night. Computer simulations confirmed that roller shades and angled blinds effectively reduce glare from skylights while maintaining sufficient light levels. An optimal control strategy is being developed to automatically adjust shading based on weather and sunlight conditions.
This document discusses using spectral corrections to better model PV performance. It presents an overview of how the spectral distribution of sunlight changes over the day and affects PV module performance. Models that account for the spectral response of modules and changing spectral irradiance can provide more accurate estimates of performance metrics like current and power output. The document shares results of analyzing spectral data that demonstrate variations in average module response under direct normal, diffuse horizontal, and global irradiance over the course of a day due to these spectral effects. Future work aims to refine estimates of air mass effects and integrate spectral modeling into broader PV modeling.
The document examines the relationship between cloud droplet effective radius and cloud top height in deep convective clouds using CloudSat data. It finds that as cloud droplet effective radius decreases, cloud top height increases, supporting the hypothesis that aerosols can invigorate deep convective clouds. However, this negative correlation is not seen for clouds with lower cloud top heights. The relationship also depends on surface temperature, with the threshold altitude for a negative correlation increasing at higher temperatures.
In today's world, natural disasters are an ever increasing presence in our lives and working environments. SolSpec has developed a unique capacity for utilizing aerial imagery to prevent and mitigate disasters.
Machine learning for evaluating climate model projectionsZachary Labe
IEEE-Student Branch- IIT Indore, Tech-Talks 2.0. Remote Presentation (Dec 2022) (Invited).
The popularity of machine learning methods, such as neural networks, is rapidly expanding in nearly all areas of science. The interest in these tools also coincides with a growing influx of big data and the need for high efficiency in solving predication problems. However, there is also some hesitancy in adopting the use of neural networks due to concerns about their reliability, reproducibility, and interpretability.
In climate science, we often consider signal-to-noise problems to help disentangle human-caused climate change from natural variability. These applications typically involve complicated relationships between different feedbacks at play in the ocean, cryosphere, land, and atmosphere. Recent work has shown that neural networks can be a promising tool for solving these types of statistical problems when combined with explainability techniques developed by the fields of computer science and image processing. Interestingly, these methods have revealed that neural networks often leverage regional patterns of climate change in order to make their predictions. In this webinar, I will share examples from climate science that use a few of these visualization methods to peer into the “black box” of neural networks, which help us to better understand their decision-making process while also learning new science. The same machine learning visualization methods can be easily adapted for a wide variety of applications and other scientific fields of study.
Using explainable machine learning for evaluating patterns of climate changeZachary Labe
22 February 2023…
Natural Sciences Group Seminar (Presentation): Using explainable machine learning for evaluating patterns of climate change, Washington State University Vancouver, USA. Remote Presentation.
References:
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464
Labe, Z.M. and E.A. Barnes (2022), Predicting slowdowns in decadal climate warming trends with explainable neural networks. Geophysical Research Letters, DOI:10.1029/2022GL098173
Exploring explainable machine learning for detecting changes in climateZachary Labe
9 February 2023…
Department of Earth, Ocean, and Atmospheric Science Colloquium (Presentation): Exploring explainable machine learning for detecting changes in climate, Florida State University, USA. Remote Presentation.
References:
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464
Labe, Z.M. and E.A. Barnes (2022), Predicting slowdowns in decadal climate warming trends with explainable neural networks. Geophysical Research Letters, DOI:10.1029/2022GL098173
Po-Chedley, S., J.T. Fasullo, N. Siler, Z.M. Labe, E.A. Barnes, C.J.W. Bonfils, and B.D. Santer (2022). Internal variability and forcing influence model-satellite differences in the rate of tropical tropospheric warming. Proceedings of the National Academy of Sciences, DOI:10.1073/pnas.2209431119
Learning new climate science by thinking creatively with machine learningZachary Labe
Presentation for: GFDL/AOS Summer Internship Lecture Series
The popularity of machine learning is rapidly growing in nearly all areas of Earth science. However, there is also some hesitancy in adopting the use of these methods due to concerns about their reliability, reproducibility, and interpretability – thus, they are often described as “black boxes.”
In this talk, I will introduce a few simple examples from climate science that leverage new visualization methods to peer into the machine learning “black box,” which help us to better understand their predictions while also learning new science. These same machine learning visualization tools can be easily adapted for a wide variety of applications and other scientific fields of study.
Explainable AI approach for evaluating climate models in the ArcticZachary Labe
27 March 2024…
IARPC Collaborations, Modelers’ Community of Practice (Presentation): Explainable AI approach for evaluating climate models in the Arctic. Remote Presentation.
References...
Labe, Z. M., & Barnes, E. A. (2022). Comparison of climate model large ensembles with observations in the Arctic using simple neural networks. Earth and Space Science, 9(7), e2022EA002348, https://doi.org/10.1029/2022EA002348
Forced climate signals with explainable AI and large ensemblesZachary Labe
1) The document discusses using artificial neural networks and explainable AI methods to analyze climate model simulations and detect climate signals related to different external forcings like greenhouse gases and aerosols.
2) A case study examines temperature trends in simulations with varying external forcings, and the neural network is able to more accurately predict observations when trained on a simulation with evolving greenhouse gases.
3) Explainable AI methods reveal the neural network relies more on patterns related to greenhouse gas forcing compared to aerosol or preindustrial forcing patterns when making predictions.
Explainable AI for identifying regional climate change patternsZachary Labe
13 January 2023…
Scientific Machine Learning Community (Presentation): Explainable AI for identifying regional climate change patterns, University of Leeds, UK. Remote Presentation.
Using explainable AI to identify key regions of climate change in GFDL SPEAR ...Zachary Labe
15 March 2023…
GFDL Lunchtime Seminar Series (Presentation): Using explainable AI to identify key regions of climate change in GFDL SPEAR large ensembles, Princeton, NJ.
References...
Labe, Z.M., E.A. Barnes, and J.W. Hurrell (2023). Identifying the regional emergence of climate patterns in the ARISE-SAI-1.5 simulations. https://doi.org/10.31223/X5394Z (submitted)
Creative machine learning approaches for climate change detectionZachary Labe
14 April 2023…
Resnick Young Investigators Symposium (Invited Presentation): Creative machine learning approaches for climate change detection, California Institute of Technology (Caltech), USA. https://resnick.caltech.edu/events/resnick-symposium/2023-symposium
References:
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464
Labe, Z.M. and E.A. Barnes (2022), Predicting slowdowns in decadal climate warming trends with explainable neural networks. Geophysical Research Letters, DOI:10.1029/2022GL098173
Po-Chedley, S., J.T. Fasullo, N. Siler, Z.M. Labe, E.A. Barnes, C.J.W. Bonfils, and B.D. Santer (2022). Internal variability and forcing influence model-satellite differences in the rate of tropical tropospheric warming. Proceedings of the National Academy of Sciences, DOI:10.1073/pnas.2209431119
Applications of machine learning for climate change and variabilityZachary Labe
23 February 2024…
Department of Environmental Sciences Seminar (Presentation): Applications of machine learning for climate change and variability, Rutgers University, New Brunswick, NJ.
References:
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464
Labe, Z.M. and E.A. Barnes (2022), Predicting slowdowns in decadal climate warming trends with explainable neural networks. Geophysical Research Letters, DOI:10.1029/2022GL098173
Labe, Z. M., Johnson, N. C., & Delworth, T. L. (2024). Changes in United States summer temperatures revealed by explainable neural networks. Earth's Future, DOI:10.1029/2023EF003981
Using explainable machine learning to evaluate climate change projectionsZachary Labe
5 October 2023…
Atmosphere and Ocean Climate Dynamics Seminar (Presentation): Using explainable machine learning to evaluate climate change projections, Yale University, New Haven, CT. Remote Presentation.
References...
Labe, Z.M., E.A. Barnes, and J.W. Hurrell (2023). Identifying the regional emergence of climate patterns in the ARISE-SAI-1.5 simulations. Environmental Research Letters, DOI:10.1088/1748-9326/acc81a, https://iopscience.iop.org/article/10.1088/1748-9326/acc81a
Explainable neural networks for evaluating patterns of climate change and var...Zachary Labe
12 March 2024…
Sharing Science – North American Webinar, Young Earth System Scientists (YESS) Community (Presentation): Explainable neural networks for evaluating patterns of climate change and variability. Remote Presentation.
References...
Labe, Z.M., E.A. Barnes, and J.W. Hurrell (2023). Identifying the regional emergence of climate patterns in the ARISE-SAI-1.5 simulations. Environmental Research Letters, DOI:10.1088/1748-9326/acc81a
An intro to explainable AI for polar climate scienceZachary Labe
26 March 2024…
GFDL Polar Climate Interest Group (Presentation): An intro to explainable AI for polar climate science, NOAA GFDL, Princeton, NJ.
References:
Labe, Z.M. and E.A. Barnes (2022), Comparison of climate model large ensembles with observations in the Arctic using simple neural networks. Earth and Space Science, DOI:10.1029/2022EA002348, https://doi.org/10.1029/2022EA002348
Labe, Z.M. and E.A. Barnes (2021), Detecting climate signals using explainable AI with single-forcing large ensembles. Journal of Advances in Modeling Earth Systems, DOI:10.1029/2021MS002464, https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2021MS002464
Climate change extremes by season in the United StatesZachary Labe
11 September 2023…
Hershey Horticulture Society (Presentation): Climate change extremes by season in the United States, Hershey, PA, USA.
References...
Eischeid, J.K., M.P. Hoerling, X.-W. Quan, A. Kumar, J. Barsugli, Z.M. Labe, K.E. Kunkel, C.J. Schreck III, D.R. Easterling, T. Zhang, J. Uehling, and X. Zhang (2023). Why has the summertime central U.S. warming hole not disappeared? Journal of Climate, DOI:10.1175/JCLI-D-22-0716.1
Labe, Z.M., T.R. Ault, and R. Zurita-Milla (2016), Identifying Anomalously Early Spring Onsets in the CESM Large Ensemble Project, R. Clim Dyn, DOI:10.1007/s00382-016-3313-2
Labe, Z.M., N.C. Johnson, and T.L Delworth (2023). Changes in United States summer temperatures revealed by explainable neural networks. Preprint. DOI: 10.22541/essoar.168987129.98069596/v1
T. Lucas Makinen x Imperial SBI WorkshopLucasMakinen1
1) The document discusses using neural networks to compress cosmological simulations into informative summaries or statistics in order to perform inference on cosmological parameters.
2) It describes using "Information Maximizing Neural Networks" which are trained to maximize the Fisher information of the summaries with respect to the parameters in order to capture the most cosmologically relevant information.
3) The document also introduces using graph neural networks to represent cosmological simulations as graphs with nodes for halos and edges for their connections, finding that the graph structure captures cosmological information in a way that is insensitive to network architecture details.
2013 ASPRS Track, Ozone Modeling for the Contiguous United States by Michael ...GIS in the Rockies
This document summarizes modeling methods for ground-level ozone concentrations in the contiguous United States. It describes four modeling methods tested: inverse distance weighting (IDW), ordinary kriging, generalized linear models (GLM), and geographically weighted regression (GWR). IDW and kriging account for spatial autocorrelation in the data. GLM and GWR use solar radiation and relative humidity as predictor variables. Kriging and GWR had the lowest errors when validated against new data points, though all models have limitations due to the characteristics and amount of input data. The document emphasizes that statistical models are abstractions of reality and should adhere to principles like parsimony.
We present the current activities of the German Climate Computing Center (DKRZ) related to the application of machine learning and deep learning in fundamental weather and climate research. We follow the Nature article "Deep learning and process understanding for data-driven Earth system science" (https://www.nature.com/articles/s41586-019-0912-1), elaborate on the hybrid model in the article "Physics-guided Neural Networks (PGNN): An Application in Lake Temperature Modeling" (https://arxiv.org/abs/1710.11431), and explain the recent application of Nvidia image inpaiting in the reconstruction of temperature missing data (Kadow et al. (2020), "Artificial Intelligence reconstructs missing Climate Information" (in review)).
What is Computer Science?
Computer Science and the Liberal Arts
The Apollo Guidance Computer
Recursive Definitions and hippopotomonstrosesquipedaliophobia
Generalizing Scientific Machine Learning and Differentiable Simulation Beyond...Chris Rackauckas
This document discusses scientific machine learning and differentiable simulation. It begins by explaining that scientific machine learning uses both data and physical knowledge to make accurate predictions with less data. It then discusses differentiable simulation and how universal differential equations can be used to replace unknown portions of models with neural networks while preserving known physical structure. Several examples are provided of applications in various domains like epidemiology, black hole detection, earthquake engineering, and chemistry. The document emphasizes that understanding the engineering principles and numerical properties of the domain is important for applying these methods stably and efficiently.
Similar to Learning new climate science by opening the machine learning black box (20)
Reexamining future projections of Arctic climate linkagesZachary Labe
10 May 2024…
Atmospheric and Oceanic Sciences Student/Postdoc Seminar (Presentation): Reexamining future projections of Arctic climate linkages, Princeton University, USA.
References...
Labe, Z.M., Y. Peings, and G. Magnusdottir (2018), Contributions of ice thickness to the atmospheric response from projected Arctic sea ice loss,
Geophysical Research Letters, DOI:10.1029/2018GL078158
Labe, Z.M., Y. Peings, and G. Magnusdottir (2019). The effect of QBO phase on the atmospheric response to projected Arctic sea ice loss in early winter, Geophysical Research Letters, DOI:10.1029/2019GL083095
Labe, Z.M., Y. Peings, and G. Magnusdottir (2020). Warm Arctic, cold Siberia pattern: role of full Arctic amplification versus sea ice loss alone, Geophysical Research Letters, DOI:10.1029/2020GL088583
Labe, Z.M., May 2020: The effects of Arctic sea-ice thickness loss and stratospheric variability on mid-latitude cold spells. University of California, Irvine. Doctoral Dissertation.
Peings, Y., Z.M. Labe, and G. Magnusdottir (2021), Are 100 ensemble members enough to capture the remote atmospheric response to +2°C Arctic sea ice loss? Journal of Climate, DOI:10.1175/JCLI-D-20-0613.1
Techniques and Considerations for Improving Accessibility in Online MediaZachary Labe
3 April 2024…
United States Association of Polar Early Career Scientists (USAPECS) IDEA Training Course (Presentation): Accessibility and disability in online spaces. Remote Presentation.
Using accessible data to communicate global climate changeZachary Labe
25 March 2024…
Climate Communication Workshop: Learn How To Make Your Research Matter (Keynote Presentation): Using accessible data to communicate global climate change, Temple University, Philadelphia, PA.
Water in a Frozen Arctic: Cross-Disciplinary PerspectivesZachary Labe
14 March 2024…
United States Association of Polar Early Career Scientists (USAPECS) Webinar (Host): Water in a Frozen Arctic: Cross-Disciplinary Perspectives. Remote Panel.
Event Page: https://www.usapecs.org/post/webinar-water-frozen-arctic
data-driven approach to identifying key regions of change associated with fut...Zachary Labe
Labe, Z.M., T.L. Delworth, N.C. Johnson, and W.F. Cooke. A data-driven approach to identifying key regions of change associated with future climate scenarios, 23rd Conference on Artificial Intelligence for Environmental Science, Baltimore, MD (Jan 2024). https://ams.confex.com/ams/104ANNUAL/meetingapp.cgi/Paper/431300
Distinguishing the regional emergence of United States summer temperatures be...Zachary Labe
Labe, Z.M., N.C. Johnson, and T.L. Delworth. Distinguishing the regional emergence of United States summer temperatures between observations and climate model large ensembles, 23rd Conference on Artificial Intelligence for Environmental Science, Baltimore, MD (Jan 2024). https://ams.confex.com/ams/104ANNUAL/meetingapp.cgi/Paper/431288
Researching and Communicating Our Changing ClimateZachary Labe
Zachary Labe is a postdoc researcher at NOAA GFDL and Princeton University who studies climate variability and change. His research uses tools like artificial intelligence and climate models to disentangle the signal of climate change from natural weather noise. He conducts field work including Arctic expeditions and uses supercomputers to run complex climate models that generate huge amounts of data.
Revisiting projections of Arctic climate change linkagesZachary Labe
16 November 2023…
Department Seminar (Presentation): Revisiting projections of Arctic climate change linkages, Tongji University, Shanghai, China. Remote Presentation.
References:
Labe, Z.M., Y. Peings, and G. Magnusdottir (2018), Contributions of ice thickness to the atmospheric response from projected Arctic sea ice loss, Geophysical Research Letters, DOI: 10.1029/2018GL078158
Labe, Z.M., Y. Peings, and G. Magnusdottir (2019). The effect of QBO phase on the atmospheric response to projected Arctic sea ice loss in early winter, Geophysical Research Letters, DOI: 10.1029/2019GL083095
Labe, Z.M., Y. Peings, and G. Magnusdottir (2020). Warm Arctic, cold Siberia pattern: role of full Arctic amplification versus sea ice loss alone, Geophysical Research Letters, DOI: 10.1029/2020GL088583
Peings, Y., Z.M. Labe, and G. Magnusdottir (2021), Are 100 ensemble members enough to capture the remote atmospheric response to +2°C Arctic sea ice loss?
Journal of Climate, DOI: 10.1175/JCLI-D-20-0613.1
Labe, Z.M. and E.A. Barnes (2022), Comparison of climate model large ensembles with observations in the Arctic using simple neural networks. Earth and Space Science, DOI: 10.1029/2022EA002348
Visualizing climate change through dataZachary Labe
18 November 2023…
NJ State Museum Planetarium (Presentation): Visualizing climate change through data, Trenton, NJ.
References...
Eischeid, J.K., M.P. Hoerling, X.-W. Quan, A. Kumar, J. Barsugli, Z.M. Labe, K.E. Kunkel, C.J. Schreck III, D.R. Easterling, T. Zhang, J. Uehling, and X. Zhang (2023). Why has the summertime central U.S. warming hole not disappeared? Journal of Climate, DOI:10.1175/JCLI-D-22-0716.1, https://journals.ametsoc.org/view/journals/clim/36/20/JCLI-D-22-0716.1.xml
Contrasting polar climate change in the past, present, and futureZachary Labe
28 September 2023…
Guest lecture for “Observing and Modeling Climate Change (EES 3506/5506)” (Presentation): Contrasting polar climate change in the past, present, and future, Temple University, Philadelphia, PA. Remote Presentation.
Guest Lecture: Our changing Arctic in the past and futureZachary Labe
22 August 2023…
Guest lecture for “Introduction to Global Climate Change (ESS 15)” (Invited): Our changing Arctic in the past and future, University of California, Irvine, CA. Remote Presentation.
References...
Delworth, T. L., Cooke, W. F., Adcroft, A., Bushuk, M., Chen, J. H., Dunne, K. A., ... & Zhao, M. (2020). SPEAR: The next generation GFDL modeling system for seasonal to multidecadal prediction and projection. Journal of Advances in Modeling Earth Systems, 12(3), e2019MS001895, https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2019MS001895
Labe, Z.M. and E.A. Barnes (2022), Comparison of climate model large ensembles with observations in the Arctic using simple neural networks. Earth and Space Science, DOI:10.1029/2022EA002348, https://doi.org/10.1029/2022EA002348
Labe, Z.M., Y. Peings, and G. Magnusdottir (2020). Warm Arctic, cold Siberia pattern: role of full Arctic amplification versus sea ice loss alone, Geophysical Research Letters, DOI:10.1029/2020GL088583, https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020GL088583
Monitoring indicators of climate change through data-driven visualizationZachary Labe
19 June 2023…
La Uni Climática - IV Edition (Presentation): Monitoring indicators of climate change through data-driven visualization. Remote Presentation.
Career pathways and research opportunities in the Earth sciencesZachary Labe
20 April 2023…
Mercer County Community College (Presentation): Career pathways and research opportunities in the Earth sciences, West Windsor Township, NJ, USA.
24 March 2023…
NBCU Fellows Academy Workshop (Presentation): Telling data-driven climate stories, The City College of New York, USA. Remote Presentation.
Evaluating and communicating Arctic climate change projectionZachary Labe
20 February 2023…
Climate Change and Agriculture Guest (Presentation): Evaluating and communicating Arctic climate change projections, Kansas State University, USA.
References...
Delworth, T. L., Cooke, W. F., Adcroft, A., Bushuk, M., Chen, J. H., Dunne, K. A., ... & Zhao, M. (2020). SPEAR: The next generation GFDL modeling system for seasonal to multidecadal prediction and projection. Journal of Advances in Modeling Earth Systems, 12(3), e2019MS001895, https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2019MS001895
Labe, Z.M. and E.A. Barnes (2022), Comparison of climate model large ensembles with observations in the Arctic using simple neural networks. Earth and Space Science, DOI:10.1029/2022EA002348, https://doi.org/10.1029/2022EA002348
Labe, Z.M., Y. Peings, and G. Magnusdottir (2020). Warm Arctic, cold Siberia pattern: role of full Arctic amplification versus sea ice loss alone, Geophysical Research Letters, DOI:10.1029/2020GL088583, https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020GL088583
Peings, Y., Cattiaux, J., Vavrus, S. J., & Magnusdottir, G. (2018). Projected squeezing of the wintertime North-Atlantic jet. Environmental Research Letters, 13(7), 074016, https://iopscience.iop.org/article/10.1088/1748-9326/aacc79/meta
ESR spectroscopy in liquid food and beverages.pptxPRIYANKA PATEL
With increasing population, people need to rely on packaged food stuffs. Packaging of food materials requires the preservation of food. There are various methods for the treatment of food to preserve them and irradiation treatment of food is one of them. It is the most common and the most harmless method for the food preservation as it does not alter the necessary micronutrients of food materials. Although irradiated food doesn’t cause any harm to the human health but still the quality assessment of food is required to provide consumers with necessary information about the food. ESR spectroscopy is the most sophisticated way to investigate the quality of the food and the free radicals induced during the processing of the food. ESR spin trapping technique is useful for the detection of highly unstable radicals in the food. The antioxidant capability of liquid food and beverages in mainly performed by spin trapping technique.
Mending Clothing to Support Sustainable Fashion_CIMaR 2024.pdfSelcen Ozturkcan
Ozturkcan, S., Berndt, A., & Angelakis, A. (2024). Mending clothing to support sustainable fashion. Presented at the 31st Annual Conference by the Consortium for International Marketing Research (CIMaR), 10-13 Jun 2024, University of Gävle, Sweden.
The cost of acquiring information by natural selectionCarl Bergstrom
This is a short talk that I gave at the Banff International Research Station workshop on Modeling and Theory in Population Biology. The idea is to try to understand how the burden of natural selection relates to the amount of information that selection puts into the genome.
It's based on the first part of this research paper:
The cost of information acquisition by natural selection
Ryan Seamus McGee, Olivia Kosterlitz, Artem Kaznatcheev, Benjamin Kerr, Carl T. Bergstrom
bioRxiv 2022.07.02.498577; doi: https://doi.org/10.1101/2022.07.02.498577
Authoring a personal GPT for your research and practice: How we created the Q...Leonel Morgado
Thematic analysis in qualitative research is a time-consuming and systematic task, typically done using teams. Team members must ground their activities on common understandings of the major concepts underlying the thematic analysis, and define criteria for its development. However, conceptual misunderstandings, equivocations, and lack of adherence to criteria are challenges to the quality and speed of this process. Given the distributed and uncertain nature of this process, we wondered if the tasks in thematic analysis could be supported by readily available artificial intelligence chatbots. Our early efforts point to potential benefits: not just saving time in the coding process but better adherence to criteria and grounding, by increasing triangulation between humans and artificial intelligence. This tutorial will provide a description and demonstration of the process we followed, as two academic researchers, to develop a custom ChatGPT to assist with qualitative coding in the thematic data analysis process of immersive learning accounts in a survey of the academic literature: QUAL-E Immersive Learning Thematic Analysis Helper. In the hands-on time, participants will try out QUAL-E and develop their ideas for their own qualitative coding ChatGPT. Participants that have the paid ChatGPT Plus subscription can create a draft of their assistants. The organizers will provide course materials and slide deck that participants will be able to utilize to continue development of their custom GPT. The paid subscription to ChatGPT Plus is not required to participate in this workshop, just for trying out personal GPTs during it.
hematic appreciation test is a psychological assessment tool used to measure an individual's appreciation and understanding of specific themes or topics. This test helps to evaluate an individual's ability to connect different ideas and concepts within a given theme, as well as their overall comprehension and interpretation skills. The results of the test can provide valuable insights into an individual's cognitive abilities, creativity, and critical thinking skills
PPT on Direct Seeded Rice presented at the three-day 'Training and Validation Workshop on Modules of Climate Smart Agriculture (CSA) Technologies in South Asia' workshop on April 22, 2024.
(June 12, 2024) Webinar: Development of PET theranostics targeting the molecu...Scintica Instrumentation
Targeting Hsp90 and its pathogen Orthologs with Tethered Inhibitors as a Diagnostic and Therapeutic Strategy for cancer and infectious diseases with Dr. Timothy Haystead.
ESA/ACT Science Coffee: Diego Blas - Gravitational wave detection with orbita...Advanced-Concepts-Team
Presentation in the Science Coffee of the Advanced Concepts Team of the European Space Agency on the 07.06.2024.
Speaker: Diego Blas (IFAE/ICREA)
Title: Gravitational wave detection with orbital motion of Moon and artificial
Abstract:
In this talk I will describe some recent ideas to find gravitational waves from supermassive black holes or of primordial origin by studying their secular effect on the orbital motion of the Moon or satellites that are laser ranged.
Current Ms word generated power point presentation covers major details about the micronuclei test. It's significance and assays to conduct it. It is used to detect the micronuclei formation inside the cells of nearly every multicellular organism. It's formation takes place during chromosomal sepration at metaphase.
Learning new climate science by opening the machine learning black box
1. LEARNING NEW CLIMATE SCIENCE BY
OPENING THE
MACHINE LEARNING BLACK BOX
@ZLabe
Zachary M. Labe
with Elizabeth A. Barnes
in the Department of Atmospheric Science
10 September 2021
Cognitive Brown Bag Series
CSU Department of Psychology
5. Computer Science
Artificial Intelligence
Machine Learning
Deep Learning
Supervised
Learning
Unsupervised
Learning
Labeled data
Classification
Regression
Unlabeled data
Clustering
Dimension reduction
6. • Do it better
• e.g., parameterizations in climate models are not
perfect, use ML to make them more accurate
• Do it faster
• e.g., code in climate models is very slow (but we
know the right answer) - use ML methods to speed
things up
• Do something new
• e.g., go looking for non-linear relationships you
didn’t know were there
Very relevant for
research: may be
slower and worse,
but can still learn
something
WHY SHOULD WE CONSIDER
MACHINE LEARNING?
9. Adapted from: Kotamarthi, R., Hayhoe, K., Mearns, L., Wuebbles, D., Jacobs, J., & Jurado, J.
(2021). Global Climate Models. In Downscaling Techniques for High-Resolution Climate
Projections: From Global Change to Local Impacts (pp. 19-39). Cambridge: Cambridge University
Press. doi:10.1017/9781108601269.003
CLIMATE MODELS
Horizontal Grid
Vertical Levels
Past/Present/Future
Fully-Coupled System
20-40 Petabytes of data
15. Today’s weather or climate
scientist is far more likely to be
debugging code written in
Python… than to be poring over
satellite images or releasing
radiosondes.
“
D. Irving| Bulletin of the American Meteorological Society| 2016
16. Machine learning for weather
IDENTIFYING SEVERE THUNDERSTORMS
Molina et al. 2021
Toms et al. 2021
CLASSIFYING PHASE OF MADDEN-JULLIAN OSCILLATION
SATELLITE DETECTION
Lee et al. 2021
DETECTING TORNADOES
McGovern et al. 2019
17. Machine learning for climate
FINDING FORECASTS OF OPPORTUNITY
Mayer and Barnes, 2021
PREDICTING CLIMATE MODES OF VARIABILITY
Gordon et al. 2021, in review
TIMING OF CLIMATE CHANGE
Barnes et al. 2019
22. E.g.,
Establish robust,
responsible AI for
severe weather
detection
Tornado Warning
Special Marine Warning
Severe Thunderstorm Warning
Flash Flood Warning
26. Artificial Neural Networks [ANN]
X1
X2
W1
W2
∑
INPUTS
NODE
Linear regression with non-linear
mapping by an “activation function”
Training of the network is merely
determining the weights “w” and
bias/offset “b"
= factivation(X1W1+ X2W2 + b)
28. Complexity and nonlinearities of the ANN allow it to learn many
different pathways of predictable behavior
Once trained, you have an array of weights and biases which can be
used for prediction on new data
INPUT
[DATA]
PREDICTION
Artificial Neural Networks [ANN]
30. ----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
TEMPERATURE
We know some metadata…
+ What year is it?
+ Where did it come from?
31. ----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
We know some metadata…
+ What year is it?
+ Where did it come from?
[e.g., Barnes et al. 2019, 2020]
[e.g., Labe and Barnes, 2021]
TIMING OF EMERGENCE
(COMBINED VARIABLES)
RESPONSES TO
EXTERNAL CLIMATE
FORCINGS
PATTERNS OF
CLIMATE INDICATORS
[e.g., Rader et al. 2021, in prep]
Surface Temperature Map Precipitation Map
+
TEMPERATURE
32. ----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
We know some metadata…
+ What year is it?
+ Where did it come from?
[e.g., Barnes et al. 2019, 2020]
[e.g., Labe and Barnes, 2021]
TIMING OF EMERGENCE
(COMBINED VARIABLES)
RESPONSES TO
EXTERNAL CLIMATE
FORCINGS
PATTERNS OF
CLIMATE INDICATORS
[e.g., Rader et al. 2021, in prep]
Surface Temperature Map Precipitation Map
+
TEMPERATURE
34. What is the annual mean temperature of Earth?
THE REAL WORLD
(Observations)
Anomaly is relative to 1951-1980
35. What is the annual mean temperature of Earth?
THE REAL WORLD
(Observations)
Let’s run a
climate model
36. What is the annual mean temperature of Earth?
THE REAL WORLD
(Observations)
Let’s run a
climate model
again
37. What is the annual mean temperature of Earth?
THE REAL WORLD
(Observations)
Let’s run a
climate model
again & again
38. What is the annual mean temperature of Earth?
THE REAL WORLD
(Observations)
CLIMATE MODEL
ENSEMBLES
39. What is the annual mean temperature of Earth?
THE REAL WORLD
(Observations)
CLIMATE MODEL
ENSEMBLES
Range of ensembles
= internal variability (noise)
Mean of ensembles
= forced response (climate change)
40. What is the annual mean temperature of Earth?
Range of ensembles
= internal variability (noise)
Mean of ensembles
= forced response (climate change)
But let’s remove
climate change…
41. What is the annual mean temperature of Earth?
Range of ensembles
= internal variability (noise)
Mean of ensembles
= forced response (climate change)
After removing the
forced response…
anomalies/noise!
42. What is the annual mean temperature of Earth?
• Increasing greenhouse gases (CO2, CH4, N2O)
• Changes in industrial aerosols (SO4, BC, OC)
• Changes in biomass burning (aerosols)
• Changes in land-use & land-cover (albedo)
43. What is the annual mean temperature of Earth?
• Increasing greenhouse gases (CO2, CH4, N2O)
• Changes in industrial aerosols (SO4, BC, OC)
• Changes in biomass burning (aerosols)
• Changes in land-use & land-cover (albedo)
Plus everything else…
(Natural/internal variability)
45. Greenhouse gases fixed to 1920 levels
All forcings (CESM-LE)
Industrial aerosols fixed to 1920 levels
[Deser et al. 2020, JCLI]
Fully-coupled CESM1.1
20 Ensemble Members
Run from 1920-2080
Observations
46. So what?
Greenhouse gases = warming
Aerosols = ?? (though mostly cooling)
What are the relative responses
between greenhouse gas
and aerosol forcing?
49. INPUT LAYER
Surface Temperature Map
ARTIFICIAL NEURAL NETWORK (ANN)
Collection of nodes (neurons)
that adjust their weights and
biases across layers in order to
learn signals for making
predictions
Learns nonlinear processes
through selected parameters
in the model
50. INPUT LAYER
HIDDEN LAYERS
OUTPUT LAYER
Surface Temperature Map
“2000-2009”
DECADE CLASS
“2070-2079”
“1920-1929”
ARTIFICIAL NEURAL NETWORK (ANN)
51. INPUT LAYER
HIDDEN LAYERS
OUTPUT LAYER
Surface Temperature Map
“2000-2009”
DECADE CLASS
“2070-2079”
“1920-1929”
BACK-PROPAGATE THROUGH NETWORK = EXPLAINABLE AI
ARTIFICIAL NEURAL NETWORK (ANN)
52. INPUT LAYER
HIDDEN LAYERS
OUTPUT LAYER
Layer-wise Relevance Propagation
Surface Temperature Map
“2000-2009”
DECADE CLASS
“2070-2079”
“1920-1929”
BACK-PROPAGATE THROUGH NETWORK = EXPLAINABLE AI
ARTIFICIAL NEURAL NETWORK (ANN)
[Barnes et al. 2020, JAMES]
[Labe and Barnes 2021, JAMES]
53. LAYER-WISE RELEVANCE PROPAGATION (LRP)
Volcano
Great White
Shark
Timber
Wolf
Image Classification LRP
https://heatmapping.org/
LRP heatmaps show regions
of “relevance” that
contribute to the neural
network’s decision-making
process for a sample
belonging to a particular
output category
Neural Network
WHY
WHY
WHY
Backpropagation – LRP
54. LAYER-WISE RELEVANCE PROPAGATION (LRP)
Volcano
Great White
Shark
Timber
Wolf
Image Classification LRP
https://heatmapping.org/
LRP heatmaps show regions
of “relevance” that
contribute to the neural
network’s decision-making
process for a sample
belonging to a particular
output category
Neural Network
WHY
WHY
WHY
Backpropagation – LRP
55. LAYER-WISE RELEVANCE PROPAGATION (LRP)
Volcano
Great White
Shark
Timber
Wolf
Image Classification LRP
https://heatmapping.org/
LRP heatmaps show regions
of “relevance” that
contribute to the neural
network’s decision-making
process for a sample
belonging to a particular
output category
Neural Network
Backpropagation – LRP
WHY
WHY
WHY
56. LAYER-WISE RELEVANCE PROPAGATION (LRP)
Image Classification LRP
https://heatmapping.org/
NOT PERFECT
Crock
Pot
Neural Network
Backpropagation – LRP
WHY
57. [Adapted from Adebayo et al., 2020]
EXPLAINABLE AI IS
NOT PERFECT
THERE ARE MANY
METHODS
58. [Adapted from Adebayo et al., 2020]
THERE ARE MANY
METHODS
EXPLAINABLE AI IS
NOT PERFECT
60. Visualizing something we already know…
Input maps of sea surface
temperatures to identify
El Niño & La Niña
Use ‘LRP’ to see how the
neural network is making
its decision
[Toms et al. 2020, JAMES]
61. OUTPUT LAYER
Layer-wise Relevance Propagation
“2000-2009”
DECADE CLASS
“2070-2079”
“1920-1929”
BACK-PROPAGATE THROUGH NETWORK = EXPLAINABLE AI
WHY?
= LRP HEAT MAPS
[Labe and Barnes 2021, JAMES]
63. Layer-wise Relevance Propagation
BACK-PROPAGATE THROUGH NETWORK = EXPLAINABLE AI
WHY?
= LRP HEAT MAPS
Find regions of “relevance”
that contribute to the
neural network’s
decision-making process
[Labe and Barnes 2021, JAMES]
64. 1960-1999: ANNUAL MEAN TEMPERATURE TRENDS
Greenhouse gases fixed
to 1920 levels
[AEROSOLS PREVAIL]
Industrial aerosols fixed
to 1920 levels
[GREENHOUSE GASES PREVAIL]
All forcings
[STANDARD CESM-LE]
DATA
65. 1960-1999: ANNUAL MEAN TEMPERATURE TRENDS
Greenhouse gases fixed
to 1920 levels
[AEROSOLS PREVAIL]
Industrial aerosols fixed
to 1920 levels
[GREENHOUSE GASES PREVAIL]
All forcings
[STANDARD CESM-LE]
DATA
66. 1960-1999: ANNUAL MEAN TEMPERATURE TRENDS
Greenhouse gases fixed
to 1920 levels
[AEROSOLS PREVAIL]
Industrial aerosols fixed
to 1920 levels
[GREENHOUSE GASES PREVAIL]
All forcings
[STANDARD CESM-LE]
DATA
67. 1960-1999: ANNUAL MEAN TEMPERATURE TRENDS
Greenhouse gases fixed
to 1920 levels
[AEROSOLS PREVAIL]
Industrial aerosols fixed
to 1920 levels
[GREENHOUSE GASES PREVAIL]
All forcings
[STANDARD CESM-LE]
DATA
68. CLIMATE MODEL DATA PREDICT THE YEAR FROM MAPS OF TEMPERATURE
AEROSOLS
PREVAIL
GREENHOUSE GASES
PREVAIL
STANDARD
CLIMATE MODEL
[Labe and Barnes 2021, JAMES]
69. OBSERVATIONS PREDICT THE YEAR FROM MAPS OF TEMPERATURE
AEROSOLS
PREVAIL
GREENHOUSE GASES
PREVAIL
STANDARD
CLIMATE MODEL
[Labe and Barnes 2021, JAMES]
70. OBSERVATIONS
SLOPES
PREDICT THE YEAR FROM MAPS OF TEMPERATURE
AEROSOLS
PREVAIL
GREENHOUSE GASES
PREVAIL
STANDARD
CLIMATE MODEL
[Labe and Barnes 2021, JAMES]
76. 1. Shuffle ensemble member and year
dimensions (bootstrap-like method)
2. Apply true labels (unshuffled years)
3. Apply same ANN architecture and LRP
4. Repeat 500x by using different
combinations of training/testing data and
initialization seeds
5. Compute 95th percentile of the distribution
of LRP at all grid points
Uncertainty in LRP
[Labe and Barnes 2021, JAMES]
77. Uncertainty in LRP
Ultimately, we are trying to
mask noise in the LRP output
Identify robust climate pattern indicators!
[Labe and Barnes 2021, JAMES]
80. Higher LRP values indicate greater relevance
for the ANN’s prediction
AVERAGED OVER 1960-2039
Aerosol-driven
Greenhouse gas-driven
All forcings
Low High
[Labe and Barnes 2021, JAMES]
84. ----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
TEMPERATURE
We know some metadata…
+ What year is it?
+ Where did it come from?
85. STANDARD EVALUATION OF
CLIMATE MODELS
Pattern correlation
RMSE
EOFs
Trends, anomalies, mean state
Climate modes of variability
87. ----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
INPUT MAPS OF TEMPERATURE, PRECIPITATION, …
88. INPUT MAPS OF TEMPERATURE, PRECIPITATION, …
NEURAL NETWORK
----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
Classification Task
89. INPUT MAPS OF TEMPERATURE, PRECIPITATION, …
NEURAL NETWORK
WHICH MODEL DID THE MAP COME FROM?
----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
Classification Task
90. INPUT MAPS OF TEMPERATURE, PRECIPITATION, …
NEURAL NETWORK
WHICH MODEL DID THE MAP COME FROM?
----ANN----
2 Hidden Layers
10 Nodes each
Ridge Regularization
Early Stopping
Classification Task
7 Classes – Large Ensemble GCMs
CMIP5 – MMLEA-SMILE Archive
16 Ensembles Per GCM
8th
class = Multi-Model Mean
97. WE CAN LEARN NEW SCIENCE
FROM EXPLAINABLE AI.
3)
98. KEY POINTS
Zachary Labe
zmlabe@rams.colostate.edu
@ZLabe
1. Machine learning is just another tool to add to our scientific workflow
2. We can use explainable AI (XAI) methods to peer into the black box of machine learning
3. We can learn new science by using XAI methods in conjunction with existing statistical tools