ISMB/ECCB 2013 Keynote Goble Results may vary: what is reproducible? why do o...Carole Goble
Keynote given by Carole Goble on 23rd July 2013 at ISMB/ECCB 2013
http://www.iscb.org/ismbeccb2013
How could we evaluate research and researchers? Reproducibility underpins the scientific method: at least in principle if not practice. The willing exchange of results and the transparent conduct of research can only be expected up to a point in a competitive environment. Contributions to science are acknowledged, but not if the credit is for data curation or software. From a bioinformatics view point, how far could our results be reproducible before the pain is just too high? Is open science a dangerous, utopian vision or a legitimate, feasible expectation? How do we move bioinformatics from one where results are post-hoc "made reproducible", to pre-hoc "born reproducible"? And why, in our computational information age, do we communicate results through fragmented, fixed documents rather than cohesive, versioned releases? I will explore these questions drawing on 20 years of experience in both the development of technical infrastructure for Life Science and the social infrastructure in which Life Science operates.
RARE and FAIR Science: Reproducibility and Research ObjectsCarole Goble
Keynote at JISC Digifest 2015 on Reproducibility and Research Objects in Scholarly Communication
Includes hidden slides
All material except maybe the IT Crowd screengrab reusable
Results Vary: The Pragmatics of Reproducibility and Research Object FrameworksCarole Goble
Keynote presentation at the iConference 2015, Newport Beach, Los Angeles, 26 March 2015.
Results Vary: The Pragmatics of Reproducibility and Research Object Frameworks
http://ischools.org/the-iconference/
BEWARE: presentation includes hidden slides AND in situ build animations - best viewed by downloading.
Keynote for Theory and Practice of Digital Libraries 2017
The theory and practice of digital libraries provides a long history of thought around how to manage knowledge ranging from collection development, to cataloging and resource description. These tools were all designed to make knowledge findable and accessible to people. Even technical progress in information retrieval and question answering are all targeted to helping answer a human’s information need.
However, increasingly demand is for data. Data that is needed not for people’s consumption but to drive machines. As an example of this demand, there has been explosive growth in job openings for Data Engineers – professionals who prepare data for machine consumption. In this talk, I overview the information needs of machine intelligence and ask the question: Are our knowledge management techniques applicable for serving this new consumer?
The Roots: Linked data and the foundations of successful Agriculture DataPaul Groth
Some thoughts on successful data for the agricultural domain. Keynote at Linked Open Data in Agriculture
MACS-G20 Workshop in Berlin, September 27th and 28th, 2017 https://www.ktbl.de/inhalte/themen/ueber-uns/projekte/macs-g20-loda/lod/
ISMB/ECCB 2013 Keynote Goble Results may vary: what is reproducible? why do o...Carole Goble
Keynote given by Carole Goble on 23rd July 2013 at ISMB/ECCB 2013
http://www.iscb.org/ismbeccb2013
How could we evaluate research and researchers? Reproducibility underpins the scientific method: at least in principle if not practice. The willing exchange of results and the transparent conduct of research can only be expected up to a point in a competitive environment. Contributions to science are acknowledged, but not if the credit is for data curation or software. From a bioinformatics view point, how far could our results be reproducible before the pain is just too high? Is open science a dangerous, utopian vision or a legitimate, feasible expectation? How do we move bioinformatics from one where results are post-hoc "made reproducible", to pre-hoc "born reproducible"? And why, in our computational information age, do we communicate results through fragmented, fixed documents rather than cohesive, versioned releases? I will explore these questions drawing on 20 years of experience in both the development of technical infrastructure for Life Science and the social infrastructure in which Life Science operates.
RARE and FAIR Science: Reproducibility and Research ObjectsCarole Goble
Keynote at JISC Digifest 2015 on Reproducibility and Research Objects in Scholarly Communication
Includes hidden slides
All material except maybe the IT Crowd screengrab reusable
Results Vary: The Pragmatics of Reproducibility and Research Object FrameworksCarole Goble
Keynote presentation at the iConference 2015, Newport Beach, Los Angeles, 26 March 2015.
Results Vary: The Pragmatics of Reproducibility and Research Object Frameworks
http://ischools.org/the-iconference/
BEWARE: presentation includes hidden slides AND in situ build animations - best viewed by downloading.
Keynote for Theory and Practice of Digital Libraries 2017
The theory and practice of digital libraries provides a long history of thought around how to manage knowledge ranging from collection development, to cataloging and resource description. These tools were all designed to make knowledge findable and accessible to people. Even technical progress in information retrieval and question answering are all targeted to helping answer a human’s information need.
However, increasingly demand is for data. Data that is needed not for people’s consumption but to drive machines. As an example of this demand, there has been explosive growth in job openings for Data Engineers – professionals who prepare data for machine consumption. In this talk, I overview the information needs of machine intelligence and ask the question: Are our knowledge management techniques applicable for serving this new consumer?
The Roots: Linked data and the foundations of successful Agriculture DataPaul Groth
Some thoughts on successful data for the agricultural domain. Keynote at Linked Open Data in Agriculture
MACS-G20 Workshop in Berlin, September 27th and 28th, 2017 https://www.ktbl.de/inhalte/themen/ueber-uns/projekte/macs-g20-loda/lod/
Combining Explicit and Latent Web Semantics for Maintaining Knowledge GraphsPaul Groth
A look at how the thinking about Web Data and the sources of semantics can help drive decisions on combining latent and explicit knowledge. Examples from Elsevier and lots of pointers to related work.
With the explosion of interest in both enhanced knowledge management and open science, the past few years have seen considerable discussion about making scientific data “FAIR” — findable, accessible, interoperable, and reusable. The problem is that most scientific datasets are not FAIR. When left to their own devices, scientists do an absolutely terrible job creating the metadata that describe the experimental datasets that make their way in online repositories. The lack of standardization makes it extremely difficult for other investigators to locate relevant datasets, to re-analyse them, and to integrate those datasets with other data. The Center for Expanded Data Annotation and Retrieval (CEDAR) has the goal of enhancing the authoring of experimental metadata to make online datasets more useful to the scientific community. The CEDAR work bench for metadata management will be presented in this webinar. CEDAR illustrates the importance of semantic technology to driving open science. It also demonstrates a means for simplifying access to scientific data sets and enhancing the reuse of the data to drive new discoveries.
The literature contains a myriad of recommendations, advice, and strictures about what data providers should do to facilitate data reuse. It can be overwhelming. Based on recent empirical work (analyzing data reuse proxies at scale, understanding data sensemaking and looking at how researchers search for data), I talk about what practices are a good place to start for helping others to reuse your data.
Being FAIR: Enabling Reproducible Data ScienceCarole Goble
Talk presented at Early Detection of Cancer Conference, OHSU, Portland, Oregon USA, 2-4 Oct 2018, http://earlydetectionresearch.com/ in the Data Science session
Dr. Dennis Wang discusses possible ways to enable ML methods to be more powerful for discovery and to reduce ambiguity within translational medicine, allowing data-informed decision-making to deliver the next generation of diagnostics and therapeutics to patients quicker, at lowered costs, and at scale.
The talk by Dr. Dennis Wang was followed by a panel discussion with Mr. Albert Wang, M. Eng., Head, IT Business Partner, Translational Research & Technologies, Bristol-Myers Squibb.
Sources of Change in Modern Knowledge Organization SystemsPaul Groth
Talk covering how knowledge graphs are making us rethink how change occurs in Knowledge Organization Systems. Based on https://arxiv.org/abs/1611.00217
Scott Edmunds talk on Big Data Publishing at the "What Bioinformaticians need to know about digital publishing beyond the PDF" workshop at ISMB 2013, July 22nd 2013
In this talk we describe how the Fourth Paradigm for Data-Intensive Research is providing a framework for us to develop tools, technologies and platforms to support actionable science. We discuss applications that take advantage of cloud computing, particularly Microsoft Azure, to realise the potential for turning data into decisions, knowledge and understanding. http://www.fourthpardigm.org and http://www.azure4research.com
Being Reproducible: SSBSS Summer School 2017Carole Goble
Lecture 2:
Being Reproducible: Models, Research Objects and R* Brouhaha
Reproducibility is a R* minefield, depending on whether you are testing for robustness (rerun), defence (repeat), certification (replicate), comparison (reproduce) or transferring between researchers (reuse). Different forms of "R" make different demands on the completeness, depth and portability of research. Sharing is another minefield raising concerns of credit and protection from sharp practices.
In practice the exchange, reuse and reproduction of scientific experiments is dependent on bundling and exchanging the experimental methods, computational codes, data, algorithms, workflows and so on along with the narrative. These "Research Objects" are not fixed, just as research is not “finished”: the codes fork, data is updated, algorithms are revised, workflows break, service updates are released. ResearchObject.org is an effort to systematically support more portable and reproducible research exchange.
In this talk I will explore these issues in more depth using the FAIRDOM Platform and its support for reproducible modelling. The talk will cover initiatives and technical issues, and raise social and cultural challenges.
Innovation applications of microphysiological systems (MPS) have been growing over the past decade, especially with respect to the use of complex human tissues for assessing safety of drug candidates – but broad industry adoption of MPS methods has not yet become a reality.
This webinar addresses some recent advances in MPS development and begins to explore the barriers to increased incorporation of MPS to improve drug safety assessment and to provide safer, more effective drugs into the clinical pipeline.
Developing data services: a tale from two Oregon universitiesAmanda Whitmire
While the generation or collection of large, complex research datasets is becoming easier and less expensive all the time, researchers often lack the knowledge and skills that are necessary to properly manage them. Having these skills is paramount in ensuring data quality, integrity, discoverability, integration, reproducibility, and reuse over time. Librarians have been preserving, managing and disseminating information for thousands of years. As scholarly research is increasingly carried out digitally, and products of research have expanded from primarily text-based manuscripts to include datasets, metadata, maps, software code etc., it is a natural expansion of scope for libraries to be involved in the stewardship of these materials as well. This kind of evolution requires that libraries bring in faculty with new skills and collaborate more intimately with researchers during the research data lifecycle, and this is exactly what is happening in academic libraries across the country. In this webinar, two researchers-turned-data-specialists, both based in academic libraries, will share their experiences and perspectives on the development of research data services at their respective institutions. Each will share their perspective on the important role that libraries can play in helping researchers manage, preserve, and share their data.
Capturing Context in Scientific Experiments: Towards Computer-Driven Sciencedgarijo
Scientists publish computational experiments in ways that do not facilitate reproducibility or reuse. Significant domain expertise, time and effort are required to understand scientific experiments and their research outputs. In order to improve this situation, mechanisms are needed to capture the exact details and the context of computational experiments. Only then, Intelligent Systems would be able help researchers understand, discover, link and reuse products of existing research.
In this presentation I will introduce my work and vision towards enabling scientists share, link, curate and reuse their computational experiments and results. In the first part of the talk, I will present my work for capturing and sharing the context of scientific experiments by using scientific workflows and machine readable representations. Thanks to this approach, experiment results are described in an unambiguous manner, have a clear trace of their creation process and include a pointer to the sources used for their generation. In the second part of the talk, I will describe examples on how the context of scientific experiments may be exploited to browse, explore and inspect research results. I will end the talk by presenting new ideas for improving and benefiting from the capture of context of scientific experiments and how to involve scientists in the process of curating and creating abstractions on available research metadata.
Drug Repurposing using Deep Learning on Knowledge GraphsDatabricks
Discovering new drugs is a lengthy and expensive process. This means that finding new uses for existing drugs can help create new treatments in less time and with less time. The difficulty is in finding these potential new uses.
How do we find these undiscovered uses for existing drugs?
We can unify the available structured and unstructured data sets into a knowledge graph. This is done by fusing the structured data sets, and performing named entity extraction on the unstructured data sets. Once this is done, we can use deep learning techniques to predict latent relationships.
In this talk we will cover:
Building the knowledge graph
Predicting latent relationships
Using the latent relationships to repurpose existing drugs
This was part of a webinar from the Materials Research Society on Machine Learning, AI, and Data-Driven Materials Development and Design. The spoken content (including Q&A) is available through MRS.
Keynote: SemSci 2017: Enabling Open Semantic Science
1st International Workshop co-located with ISWC 2017, October 2017, Vienna, Austria,
https://semsci.github.io/semSci2017/
Abstract
We have all grown up with the research article and article collections (let’s call them libraries) as the prime means of scientific discourse. But research output is more than just the rhetorical narrative. The experimental methods, computational codes, data, algorithms, workflows, Standard Operating Procedures, samples and so on are the objects of research that enable reuse and reproduction of scientific experiments, and they too need to be examined and exchanged as research knowledge.
We can think of “Research Objects” as different types and as packages all the components of an investigation. If we stop thinking of publishing papers and start thinking of releasing Research Objects (software), then scholar exchange is a new game: ROs and their content evolve; they are multi-authored and their authorship evolves; they are a mix of virtual and embedded, and so on.
But first, some baby steps before we get carried away with a new vision of scholarly communication. Many journals (e.g. eLife, F1000, Elsevier) are just figuring out how to package together the supplementary materials of a paper. Data catalogues are figuring out how to virtually package multiple datasets scattered across many repositories to keep the integrated experimental context.
Research Objects [1] (http://researchobject.org/) is a framework by which the many, nested and contributed components of research can be packaged together in a systematic way, and their context, provenance and relationships richly described. The brave new world of containerisation provides the containers and Linked Data provides the metadata framework for the container manifest construction and profiles. It’s not just theory, but also in practice with examples in Systems Biology modelling, Bioinformatics computational workflows, and Health Informatics data exchange. I’ll talk about why and how we got here, the framework and examples, and what we need to do.
[1] Sean Bechhofer, Iain Buchan, David De Roure, Paolo Missier, John Ainsworth, Jiten Bhagat, Philip Couch, Don Cruickshank, Mark Delderfield, Ian Dunlop, Matthew Gamble, Danius Michaelides, Stuart Owen, David Newman, Shoaib Sufi, Carole Goble, Why linked data is not enough for scientists, In Future Generation Computer Systems, Volume 29, Issue 2, 2013, Pages 599-611, ISSN 0167-739X, https://doi.org/10.1016/j.future.2011.08.004
This presentation was provided by Sebastian Kohlmeier of The Allen Institute for AI (AI2), during the NISO event "Transforming Search: What the Information Community Can and Should Build." The virtual conference was held on August 26, 2020.
Open & reproducible research - What can we do in practice?Felix Z. Hoffmann
Talk on my project within the Open Science Fellowship program, held at the Bordeaux Neurocampus on April 2018.
Note: For working videos, please refer to the GitHub source code http://bit.ly/bx18s
Combining Explicit and Latent Web Semantics for Maintaining Knowledge GraphsPaul Groth
A look at how the thinking about Web Data and the sources of semantics can help drive decisions on combining latent and explicit knowledge. Examples from Elsevier and lots of pointers to related work.
With the explosion of interest in both enhanced knowledge management and open science, the past few years have seen considerable discussion about making scientific data “FAIR” — findable, accessible, interoperable, and reusable. The problem is that most scientific datasets are not FAIR. When left to their own devices, scientists do an absolutely terrible job creating the metadata that describe the experimental datasets that make their way in online repositories. The lack of standardization makes it extremely difficult for other investigators to locate relevant datasets, to re-analyse them, and to integrate those datasets with other data. The Center for Expanded Data Annotation and Retrieval (CEDAR) has the goal of enhancing the authoring of experimental metadata to make online datasets more useful to the scientific community. The CEDAR work bench for metadata management will be presented in this webinar. CEDAR illustrates the importance of semantic technology to driving open science. It also demonstrates a means for simplifying access to scientific data sets and enhancing the reuse of the data to drive new discoveries.
The literature contains a myriad of recommendations, advice, and strictures about what data providers should do to facilitate data reuse. It can be overwhelming. Based on recent empirical work (analyzing data reuse proxies at scale, understanding data sensemaking and looking at how researchers search for data), I talk about what practices are a good place to start for helping others to reuse your data.
Being FAIR: Enabling Reproducible Data ScienceCarole Goble
Talk presented at Early Detection of Cancer Conference, OHSU, Portland, Oregon USA, 2-4 Oct 2018, http://earlydetectionresearch.com/ in the Data Science session
Dr. Dennis Wang discusses possible ways to enable ML methods to be more powerful for discovery and to reduce ambiguity within translational medicine, allowing data-informed decision-making to deliver the next generation of diagnostics and therapeutics to patients quicker, at lowered costs, and at scale.
The talk by Dr. Dennis Wang was followed by a panel discussion with Mr. Albert Wang, M. Eng., Head, IT Business Partner, Translational Research & Technologies, Bristol-Myers Squibb.
Sources of Change in Modern Knowledge Organization SystemsPaul Groth
Talk covering how knowledge graphs are making us rethink how change occurs in Knowledge Organization Systems. Based on https://arxiv.org/abs/1611.00217
Scott Edmunds talk on Big Data Publishing at the "What Bioinformaticians need to know about digital publishing beyond the PDF" workshop at ISMB 2013, July 22nd 2013
In this talk we describe how the Fourth Paradigm for Data-Intensive Research is providing a framework for us to develop tools, technologies and platforms to support actionable science. We discuss applications that take advantage of cloud computing, particularly Microsoft Azure, to realise the potential for turning data into decisions, knowledge and understanding. http://www.fourthpardigm.org and http://www.azure4research.com
Being Reproducible: SSBSS Summer School 2017Carole Goble
Lecture 2:
Being Reproducible: Models, Research Objects and R* Brouhaha
Reproducibility is a R* minefield, depending on whether you are testing for robustness (rerun), defence (repeat), certification (replicate), comparison (reproduce) or transferring between researchers (reuse). Different forms of "R" make different demands on the completeness, depth and portability of research. Sharing is another minefield raising concerns of credit and protection from sharp practices.
In practice the exchange, reuse and reproduction of scientific experiments is dependent on bundling and exchanging the experimental methods, computational codes, data, algorithms, workflows and so on along with the narrative. These "Research Objects" are not fixed, just as research is not “finished”: the codes fork, data is updated, algorithms are revised, workflows break, service updates are released. ResearchObject.org is an effort to systematically support more portable and reproducible research exchange.
In this talk I will explore these issues in more depth using the FAIRDOM Platform and its support for reproducible modelling. The talk will cover initiatives and technical issues, and raise social and cultural challenges.
Innovation applications of microphysiological systems (MPS) have been growing over the past decade, especially with respect to the use of complex human tissues for assessing safety of drug candidates – but broad industry adoption of MPS methods has not yet become a reality.
This webinar addresses some recent advances in MPS development and begins to explore the barriers to increased incorporation of MPS to improve drug safety assessment and to provide safer, more effective drugs into the clinical pipeline.
Developing data services: a tale from two Oregon universitiesAmanda Whitmire
While the generation or collection of large, complex research datasets is becoming easier and less expensive all the time, researchers often lack the knowledge and skills that are necessary to properly manage them. Having these skills is paramount in ensuring data quality, integrity, discoverability, integration, reproducibility, and reuse over time. Librarians have been preserving, managing and disseminating information for thousands of years. As scholarly research is increasingly carried out digitally, and products of research have expanded from primarily text-based manuscripts to include datasets, metadata, maps, software code etc., it is a natural expansion of scope for libraries to be involved in the stewardship of these materials as well. This kind of evolution requires that libraries bring in faculty with new skills and collaborate more intimately with researchers during the research data lifecycle, and this is exactly what is happening in academic libraries across the country. In this webinar, two researchers-turned-data-specialists, both based in academic libraries, will share their experiences and perspectives on the development of research data services at their respective institutions. Each will share their perspective on the important role that libraries can play in helping researchers manage, preserve, and share their data.
Capturing Context in Scientific Experiments: Towards Computer-Driven Sciencedgarijo
Scientists publish computational experiments in ways that do not facilitate reproducibility or reuse. Significant domain expertise, time and effort are required to understand scientific experiments and their research outputs. In order to improve this situation, mechanisms are needed to capture the exact details and the context of computational experiments. Only then, Intelligent Systems would be able help researchers understand, discover, link and reuse products of existing research.
In this presentation I will introduce my work and vision towards enabling scientists share, link, curate and reuse their computational experiments and results. In the first part of the talk, I will present my work for capturing and sharing the context of scientific experiments by using scientific workflows and machine readable representations. Thanks to this approach, experiment results are described in an unambiguous manner, have a clear trace of their creation process and include a pointer to the sources used for their generation. In the second part of the talk, I will describe examples on how the context of scientific experiments may be exploited to browse, explore and inspect research results. I will end the talk by presenting new ideas for improving and benefiting from the capture of context of scientific experiments and how to involve scientists in the process of curating and creating abstractions on available research metadata.
Drug Repurposing using Deep Learning on Knowledge GraphsDatabricks
Discovering new drugs is a lengthy and expensive process. This means that finding new uses for existing drugs can help create new treatments in less time and with less time. The difficulty is in finding these potential new uses.
How do we find these undiscovered uses for existing drugs?
We can unify the available structured and unstructured data sets into a knowledge graph. This is done by fusing the structured data sets, and performing named entity extraction on the unstructured data sets. Once this is done, we can use deep learning techniques to predict latent relationships.
In this talk we will cover:
Building the knowledge graph
Predicting latent relationships
Using the latent relationships to repurpose existing drugs
This was part of a webinar from the Materials Research Society on Machine Learning, AI, and Data-Driven Materials Development and Design. The spoken content (including Q&A) is available through MRS.
Keynote: SemSci 2017: Enabling Open Semantic Science
1st International Workshop co-located with ISWC 2017, October 2017, Vienna, Austria,
https://semsci.github.io/semSci2017/
Abstract
We have all grown up with the research article and article collections (let’s call them libraries) as the prime means of scientific discourse. But research output is more than just the rhetorical narrative. The experimental methods, computational codes, data, algorithms, workflows, Standard Operating Procedures, samples and so on are the objects of research that enable reuse and reproduction of scientific experiments, and they too need to be examined and exchanged as research knowledge.
We can think of “Research Objects” as different types and as packages all the components of an investigation. If we stop thinking of publishing papers and start thinking of releasing Research Objects (software), then scholar exchange is a new game: ROs and their content evolve; they are multi-authored and their authorship evolves; they are a mix of virtual and embedded, and so on.
But first, some baby steps before we get carried away with a new vision of scholarly communication. Many journals (e.g. eLife, F1000, Elsevier) are just figuring out how to package together the supplementary materials of a paper. Data catalogues are figuring out how to virtually package multiple datasets scattered across many repositories to keep the integrated experimental context.
Research Objects [1] (http://researchobject.org/) is a framework by which the many, nested and contributed components of research can be packaged together in a systematic way, and their context, provenance and relationships richly described. The brave new world of containerisation provides the containers and Linked Data provides the metadata framework for the container manifest construction and profiles. It’s not just theory, but also in practice with examples in Systems Biology modelling, Bioinformatics computational workflows, and Health Informatics data exchange. I’ll talk about why and how we got here, the framework and examples, and what we need to do.
[1] Sean Bechhofer, Iain Buchan, David De Roure, Paolo Missier, John Ainsworth, Jiten Bhagat, Philip Couch, Don Cruickshank, Mark Delderfield, Ian Dunlop, Matthew Gamble, Danius Michaelides, Stuart Owen, David Newman, Shoaib Sufi, Carole Goble, Why linked data is not enough for scientists, In Future Generation Computer Systems, Volume 29, Issue 2, 2013, Pages 599-611, ISSN 0167-739X, https://doi.org/10.1016/j.future.2011.08.004
This presentation was provided by Sebastian Kohlmeier of The Allen Institute for AI (AI2), during the NISO event "Transforming Search: What the Information Community Can and Should Build." The virtual conference was held on August 26, 2020.
Open & reproducible research - What can we do in practice?Felix Z. Hoffmann
Talk on my project within the Open Science Fellowship program, held at the Bordeaux Neurocampus on April 2018.
Note: For working videos, please refer to the GitHub source code http://bit.ly/bx18s
"Reproducibility from the Informatics Perspective"Micah Altman
Dr. Altman will provide expert comment on the need for informatics modeling as part of the National Academies workshop: Statistical Challenges in Assessing and Fostering the Reproducibility of Scientific Results
This workshop focuses on the topic of addressing statistical challenges in assessing and fostering the reproducibility of scientific results by examining three issues from a statistical perspective: the extent of reproducibility, the causes of reproducibility failures, and potential remedies.
The scientific and economic value of research data is enormous. To ensure successful subsequent usage, the scientific community needs efficient access to data, the data has to be reliable and persistent, and the quality of the data has to be proved.
One solution to these preconditions is to apply the techniques of today’s scientific publishing to research data. Besides its publication in a data repository together with some metadata, the data should undergo a transparent public peer-review using a publication platform.
The presentation discusses two approaches. On the one hand, the data can be the basis for a research article and undergoes a review parallel to the review of the manuscript. The data is then a reviewed supplement to a scientific publication. On the other hand, the data itself can be the subject of a publication whose quality is then assured by peers.
The presentation provides practical experience, especially with the latter strategy, realized through an established open access journal.
Abstract: - Data Mining is a process used to extract the usable data from a larger set of any raw data. It involves
analyzing data patterns in large batches of data using one or more software. R is a programming language for the
purpose of statistical computations and data analysis. The R language is widely used by the data miners and
statisticians on high dimensional pattern extraction. R is freely available under the GNU General Public Licenses
and the source code is written in FORTAN, C and R.It is a GNU project. The pre-compiled binary versions are
freely available for various flavours of operating system. R is basically command line interface (CLI) and various
GUI interfaces are also available now a day. In this articlefocuses the concept of R like; getting data into and out of
R and packages related to data mining and data visualization.
This is my class project using UCI Mashable dataset to determine what constitutes popular news. In this project, I used (1) multiple regression and model building and (2) PCA and factor analysis.
Data Analytics Tools: SAS and R
Similar to Reproducible research: First steps. (20)
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Subhajit Sahu
Abstract — Levelwise PageRank is an alternative method of PageRank computation which decomposes the input graph into a directed acyclic block-graph of strongly connected components, and processes them in topological order, one level at a time. This enables calculation for ranks in a distributed fashion without per-iteration communication, unlike the standard method where all vertices are processed in each iteration. It however comes with a precondition of the absence of dead ends in the input graph. Here, the native non-distributed performance of Levelwise PageRank was compared against Monolithic PageRank on a CPU as well as a GPU. To ensure a fair comparison, Monolithic PageRank was also performed on a graph where vertices were split by components. Results indicate that Levelwise PageRank is about as fast as Monolithic PageRank on the CPU, but quite a bit slower on the GPU. Slowdown on the GPU is likely caused by a large submission of small workloads, and expected to be non-issue when the computation is performed on massive graphs.
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Data and AI
Discussion on Vector Databases, Unstructured Data and AI
https://www.meetup.com/unstructured-data-meetup-new-york/
This meetup is for people working in unstructured data. Speakers will come present about related topics such as vector databases, LLMs, and managing data at scale. The intended audience of this group includes roles like machine learning engineers, data scientists, data engineers, software engineers, and PMs.This meetup was formerly Milvus Meetup, and is sponsored by Zilliz maintainers of Milvus.
Adjusting primitives for graph : SHORT REPORT / NOTESSubhajit Sahu
Graph algorithms, like PageRank Compressed Sparse Row (CSR) is an adjacency-list based graph representation that is
Multiply with different modes (map)
1. Performance of sequential execution based vs OpenMP based vector multiply.
2. Comparing various launch configs for CUDA based vector multiply.
Sum with different storage types (reduce)
1. Performance of vector element sum using float vs bfloat16 as the storage type.
Sum with different modes (reduce)
1. Performance of sequential execution based vs OpenMP based vector element sum.
2. Performance of memcpy vs in-place based CUDA based vector element sum.
3. Comparing various launch configs for CUDA based vector element sum (memcpy).
4. Comparing various launch configs for CUDA based vector element sum (in-place).
Sum with in-place strategies of CUDA mode (reduce)
1. Comparing various launch configs for CUDA based vector element sum (in-place).
2. Credibility turns on the success or failure
of attempts to reproduce findings.
Kenneth Rogoff &
Carmen Reinhart
In economic models
• coding errors
• selective exclusion of available data
• unconventional weighting of
summary statistics
Thomas Herdon, Michael Ash, & Robert Pollin (2013). Does high public debt
consistently stifle economic growth? A critique of Reinhart and Rogoff. Working
paper series 322. Political Economy Research Institute, U Mass Amherst.
3. Credibility turns on the success or failure
of attempts to reproduce findings.
Jason deBruyn (Jan 23, 2015) Trial involving disgraced scientist and
bunk Duke research to begin Monday. Triangle Business Journal.
In cancer therapy models
• data falsification
• retracted journal articles
• terminated clinical trials
• civil suit by patients
Anil Potti
4. Credibility turns on the success or failure
of attempts to reproduce findings.
1000 years of temperature variation: the
”hockey stick” graph by Michael Mann
In climate science models
• flawed research methods
• evasion of FOIA requests
• leaked emails
• media hype
Freed Pearce (2010-02-09) Climate change debate overheated
after sceptic grasped 'hockey stick‘. The Guardian.
5. “Computational science today faces a credibility crisis.”
Victoria Stodden, UIUC
Without access to the code and
data that underlie scientific
discoveries, published findings
are all but impossible to verify.
6. What can reproducible research do for you?
Your closest collaborator
is you six months ago,
but you don't reply to emails.
Paul Wilson
Engineering Physics
UW–Madison
10. Principle 1.
Blend computing, results, and narrative.
<<>>=
hist(co2)
@
Render the text and
code outputs.
Report title
Introduction.
Some narrative.
Discuss result.
More narrative.