this presentation was prepared for a minisymposium on the occasion of PhD defence of Niels Rutten June 14 2017 at Wageningen University, with the thesis entitled “The utility of sensor technology to support reproductive management on dairy farms”. The public defence of his thesis was a good reason to share knowledge about current sensor research in the dairy farming industry
How can banks maximise the value of their customer data?Ben Gilchriest
Almost all banks say that being customer centric is important to them and yet only a small proportion of customers believe that their banks really understand their needs and wants well enough (only 37%). This may be surprising given how much data banks have on their customers - a figure that has only been increasing over the past few years as more and more interactions become digitized. Add to this new sources of data which are available now on preferences,via social media, and increasingly available on location and physiology (see; http://bengilchriest.tumblr.com for more on this)....and the opportunity for better customer understanding becomes huge.
With a 90% of banks citing "big data" as key to long term success, where's the disconnecting coming from? In this study it's clear that the main challenge is that data is not sufficiently well pooled to realise the benefits of cross-referencing to gain insight. Coupled with the fact that not enough time is spent on analysis and the gap between the intent and customer's view becomes clearer.
So what can banks do about this? This paper describes some of the key challenges, which may be familiar to you, and some insights into how to scale up to the next level of customer analytics.
It includes a high level tool to assess your big data maturity.
Large Language Models, No-Code, and Responsible AI - Trends in Applied NLP in...David Talby
An April 2023 presentation to the AMIA working group on natural language processing. The talk focuses on three current trends in NLP and how they apply in healthcare: Large language models, No-code, and Responsible AI.
How can banks maximise the value of their customer data?Ben Gilchriest
Almost all banks say that being customer centric is important to them and yet only a small proportion of customers believe that their banks really understand their needs and wants well enough (only 37%). This may be surprising given how much data banks have on their customers - a figure that has only been increasing over the past few years as more and more interactions become digitized. Add to this new sources of data which are available now on preferences,via social media, and increasingly available on location and physiology (see; http://bengilchriest.tumblr.com for more on this)....and the opportunity for better customer understanding becomes huge.
With a 90% of banks citing "big data" as key to long term success, where's the disconnecting coming from? In this study it's clear that the main challenge is that data is not sufficiently well pooled to realise the benefits of cross-referencing to gain insight. Coupled with the fact that not enough time is spent on analysis and the gap between the intent and customer's view becomes clearer.
So what can banks do about this? This paper describes some of the key challenges, which may be familiar to you, and some insights into how to scale up to the next level of customer analytics.
It includes a high level tool to assess your big data maturity.
Large Language Models, No-Code, and Responsible AI - Trends in Applied NLP in...David Talby
An April 2023 presentation to the AMIA working group on natural language processing. The talk focuses on three current trends in NLP and how they apply in healthcare: Large language models, No-code, and Responsible AI.
Generative AI: Past, Present, and Future – A Practitioner's PerspectiveHuahai Yang
Generative AI: Past, Present, and Future – A Practitioner's Perspective
As the academic realm grapples with the profound implications of generative AI
and related applications like ChatGPT, I will present a grounded view from my
experience as a practitioner. Starting with the origins of neural networks in
the fields of logic, psychology, and computer science, I trace its history and
align it within the wider context of the pursuit of artificial intelligence.
This perspective will also draw parallels with historical developments in
psychology. Against this backdrop, I chart a proposed trajectory for the future.
Finally, I provide actionable insights for both academics and enterprising
individuals in the field.
Generative AI in Healthcare Market - Copy - Copy.pptxGayatriGadhave1
Generative AI holds significant promise in healthcare, there are also challenges related to data privacy, model interpretability, and regulatory compliance that need to be addressed. Ethical considerations and thorough validation processes are crucial to ensure the responsible and safe application of generative AI techniques in healthcare.
Enabling a Data Mesh Architecture with Data VirtualizationDenodo
Watch full webinar here: https://bit.ly/3rwWhyv
The Data Mesh architectural design was first proposed in 2019 by Zhamak Dehghani, principal technology consultant at Thoughtworks, a technology company that is closely associated with the development of distributed agile methodology. A data mesh is a distributed, de-centralized data infrastructure in which multiple autonomous domains manage and expose their own data, called “data products,” to the rest of the organization.
Organizations leverage data mesh architecture when they experience shortcomings in highly centralized architectures, such as the lack domain-specific expertise in data teams, the inflexibility of centralized data repositories in meeting the specific needs of different departments within large organizations, and the slow nature of centralized data infrastructures in provisioning data and responding to changes.
In this session, Pablo Alvarez, Global Director of Product Management at Denodo, explains how data virtualization is your best bet for implementing an effective data mesh architecture.
You will learn:
- How data mesh architecture not only enables better performance and agility, but also self-service data access
- The requirements for “data products” in the data mesh world, and how data virtualization supports them
- How data virtualization enables domains in a data mesh to be truly autonomous
- Why a data lake is not automatically a data mesh
- How to implement a simple, functional data mesh architecture using data virtualization
A Framework for Navigating Generative Artificial Intelligence for EnterpriseRocketSource
Generative AI has dominated the headlines recently, which has caused many enterprises to put a full stop to implementing this technology until they can understand what’s behind the glitz and glamour. What if we shifted the conversation? What if the focus became a fresh, incremental approach to embracing the opportunities with generative artificial intelligence to keep organizations moving upward on the S Curve of Growth?
Brands stay relevant and solve complex problems by testing the barometer for one thing — will a new strategy, tool, or piece of technology improve humanity?
Human connections are more vital than using shiny new tools or technology. As your teams work to steer clear of the temptation to do what everyone else is doing in uniform, this post will highlight how to stand out, compete, and do so with less risk in today’s world of generative AI overload.
AI Data Acquisition and Governance: Considerations for SuccessDatabricks
data pipeline, governance, and for growth and updating models regularly needs to be part of the AI strategy from the outset.
This session will cover:
Defining AI governance: What this means and how definitions of subjects like ethics and effectiveness can differ between organizations.
Data governance: Companies must rely on an AI governance program to ensure only high-quality, unbiased and consistent data are used in training.
AI is a growing necessity for enterprises / businesses; it provides an avenue for scaling quickly and efficiently.
Best practices / implementation: how to implement AI that meets the requirements of the organization’s defined sets of governances.
Planning the data pipeline and growing/updating the models: AI is not static in the real world; models must be frequently updated to maintain relevance and accuracy.
3 key takeaways or attendee benefits of the session:
Understand how to assess your organization’s need for AI; how to identify the opportune areas for transforming processes, interactions, scaling, cost.
How to start the implementation process. Defining data and AI governance and how to build the training data pipeline within that framework.
Best practices for maintaining AI; how to use data to evaluate models and continuously iterate on them to reflect the real world.
Our report will provide a look into the technology landscape of the future, including:
- Importance of AI in enabling innovation
- Catalysts of future innovations
- Top technology trends in 2023-2024
- Main benefits of AI adoption
- Steps to prepare for future disruptions.
Download your free copy now and implement the key findings to improve your business.
A Reference Architecture for Digital Health: The Health Catalyst Data Operati...Health Catalyst
There are essentially four strategic options to address the enterprise data platform requirements of today’s healthcare systems: (1) build your own, (2) buy from EHR vendors, (3) look to a Silicon Valley high-tech startup, and (4) partner with Health Catalyst or a handful of similar companies.
In this webinar, Health Catalyst’s CTO, Dale Sanders, comments on all four approaches, hoping to help you to assess your organization’s strategy against the options and vendors in each category.
It’s been exactly three years since Health Catalyst embarked on a major investment in its next-generation technology, the Data Operating System (DOS™) and its applications. This webinar is an update on the progress, less about marketing the technology, but rather offering DOS as a reference architecture that can support analytics, AI, text processing, data-first application development, and interoperability, as an all-in-one agile cost-savings architecture.
In addition to the successes, Dale comments on the challenges that Health Catalyst has faced under a very ambitious DOS development plan. In its current state, DOS has made some significant improvements to overcome early mistakes, and is now a very solid enterprise data platform. In the interests of industry-wide learning, Sanders will talk transparently about those mistakes and how those learnings are being applied to the DOS platform, positioning it to evolve gracefully over the next 25 years.
View the webinar to learn how the DOS reference architecture:
- Helps manage the 2,000+ compulsory measures in US healthcare
- Enables applications as varied as a real-time patient safety surveillance system, and an activity-based costing system in one platform
- Can ingest data of any type or velocity from over 300 healthcare source systems and growing
- Bundles tools, applications, and analytics that would cost 3-6x more to build on your own
- Compares to EHR vendors as an option to serve as an enterprise data and analytics platform
- Is a performant, sustainable, and maintainable platform for deploying AI models in the natural flow of the healthcare data pipeline
- Provides curated data content and models while still allowing for the agility of a late binding design option
- Functions as a reference architecture that all healthcare organizations and vendors will ultimately have to build in their pursuit of digital health
Knowledge graphs ilaria maresi the hyve 23apr2020Pistoia Alliance
Data for drug discovery and healthcare is often trapped in silos which hampers effective interpretation and reuse. To remedy this, such data needs to be linked both internally and to external sources to make a FAIR data landscape which can power semantic models and knowledge graphs.
How to Use Geospatial Data to Identify CPG Demnd HotspotsCARTO
The combination of new location data streams and spatial data science techniques opens up a new array of opportunities for CPG data and marketing professionals seeking where to prioritize in terms of ramping up distribution and identifying POS (points of sale).
Natural Language Understanding in HealthcareDavid Talby
The ability of software to reason, answer questions and intelligently converse about clinical notes, patient stories or biomedical papers has risen dramatically in the past few years.
This talk covers state of the art natural language processing, deep learning, and machine learning libraries in this space. We'll share benchmarks from industry & research projects on use cases such as clinical data abstraction, patient risk prediction, named entity recognition & resolution, and negation scope detection.
New technologies such as the Internet of Things and Cloud Computing are expected to leverage the current
trend of Smart Farming, introducing more sensors, robots and artificial intelligence, encompassed by the
phenomenon of Big Data.
This presentation will give a quick insight into the state-of-the-art of Big Data applications in Smart Farming
and identify the related challenges that have to be addressed. It shows that the scope of Big Data
applications in Smart Farming goes beyond the farm; it is influencing the entire food supply chain. Big data
are being used to provide predictive insights in farming operations, drive real-time operational decisions, and
redesign business processes for game-changing business models.
It is expected that Big Data will cause major shifts in roles and power relations among different players in
current food supply chain networks. The landscape of stakeholders exhibits an interesting game between
powerful tech companies, venture capitalists and often small startups and new entrants. At the same time
there are several public institutions that publish open data, under the condition that the privacy of persons
must be guaranteed. The future of Smart Farming may unravel in a continuum of two extreme scenarios: 1)
closed, proprietary systems or 2) open, collaborative systems.
The development of data and application infrastructures (platforms and standards) and their institutional
embedment will play a crucial role in the battle between these scenarios. A major challenge is therefore to
cope with governance issues and define suitable business models for data sharing in different supply chain
scenarios.
Generative AI: Past, Present, and Future – A Practitioner's PerspectiveHuahai Yang
Generative AI: Past, Present, and Future – A Practitioner's Perspective
As the academic realm grapples with the profound implications of generative AI
and related applications like ChatGPT, I will present a grounded view from my
experience as a practitioner. Starting with the origins of neural networks in
the fields of logic, psychology, and computer science, I trace its history and
align it within the wider context of the pursuit of artificial intelligence.
This perspective will also draw parallels with historical developments in
psychology. Against this backdrop, I chart a proposed trajectory for the future.
Finally, I provide actionable insights for both academics and enterprising
individuals in the field.
Generative AI in Healthcare Market - Copy - Copy.pptxGayatriGadhave1
Generative AI holds significant promise in healthcare, there are also challenges related to data privacy, model interpretability, and regulatory compliance that need to be addressed. Ethical considerations and thorough validation processes are crucial to ensure the responsible and safe application of generative AI techniques in healthcare.
Enabling a Data Mesh Architecture with Data VirtualizationDenodo
Watch full webinar here: https://bit.ly/3rwWhyv
The Data Mesh architectural design was first proposed in 2019 by Zhamak Dehghani, principal technology consultant at Thoughtworks, a technology company that is closely associated with the development of distributed agile methodology. A data mesh is a distributed, de-centralized data infrastructure in which multiple autonomous domains manage and expose their own data, called “data products,” to the rest of the organization.
Organizations leverage data mesh architecture when they experience shortcomings in highly centralized architectures, such as the lack domain-specific expertise in data teams, the inflexibility of centralized data repositories in meeting the specific needs of different departments within large organizations, and the slow nature of centralized data infrastructures in provisioning data and responding to changes.
In this session, Pablo Alvarez, Global Director of Product Management at Denodo, explains how data virtualization is your best bet for implementing an effective data mesh architecture.
You will learn:
- How data mesh architecture not only enables better performance and agility, but also self-service data access
- The requirements for “data products” in the data mesh world, and how data virtualization supports them
- How data virtualization enables domains in a data mesh to be truly autonomous
- Why a data lake is not automatically a data mesh
- How to implement a simple, functional data mesh architecture using data virtualization
A Framework for Navigating Generative Artificial Intelligence for EnterpriseRocketSource
Generative AI has dominated the headlines recently, which has caused many enterprises to put a full stop to implementing this technology until they can understand what’s behind the glitz and glamour. What if we shifted the conversation? What if the focus became a fresh, incremental approach to embracing the opportunities with generative artificial intelligence to keep organizations moving upward on the S Curve of Growth?
Brands stay relevant and solve complex problems by testing the barometer for one thing — will a new strategy, tool, or piece of technology improve humanity?
Human connections are more vital than using shiny new tools or technology. As your teams work to steer clear of the temptation to do what everyone else is doing in uniform, this post will highlight how to stand out, compete, and do so with less risk in today’s world of generative AI overload.
AI Data Acquisition and Governance: Considerations for SuccessDatabricks
data pipeline, governance, and for growth and updating models regularly needs to be part of the AI strategy from the outset.
This session will cover:
Defining AI governance: What this means and how definitions of subjects like ethics and effectiveness can differ between organizations.
Data governance: Companies must rely on an AI governance program to ensure only high-quality, unbiased and consistent data are used in training.
AI is a growing necessity for enterprises / businesses; it provides an avenue for scaling quickly and efficiently.
Best practices / implementation: how to implement AI that meets the requirements of the organization’s defined sets of governances.
Planning the data pipeline and growing/updating the models: AI is not static in the real world; models must be frequently updated to maintain relevance and accuracy.
3 key takeaways or attendee benefits of the session:
Understand how to assess your organization’s need for AI; how to identify the opportune areas for transforming processes, interactions, scaling, cost.
How to start the implementation process. Defining data and AI governance and how to build the training data pipeline within that framework.
Best practices for maintaining AI; how to use data to evaluate models and continuously iterate on them to reflect the real world.
Our report will provide a look into the technology landscape of the future, including:
- Importance of AI in enabling innovation
- Catalysts of future innovations
- Top technology trends in 2023-2024
- Main benefits of AI adoption
- Steps to prepare for future disruptions.
Download your free copy now and implement the key findings to improve your business.
A Reference Architecture for Digital Health: The Health Catalyst Data Operati...Health Catalyst
There are essentially four strategic options to address the enterprise data platform requirements of today’s healthcare systems: (1) build your own, (2) buy from EHR vendors, (3) look to a Silicon Valley high-tech startup, and (4) partner with Health Catalyst or a handful of similar companies.
In this webinar, Health Catalyst’s CTO, Dale Sanders, comments on all four approaches, hoping to help you to assess your organization’s strategy against the options and vendors in each category.
It’s been exactly three years since Health Catalyst embarked on a major investment in its next-generation technology, the Data Operating System (DOS™) and its applications. This webinar is an update on the progress, less about marketing the technology, but rather offering DOS as a reference architecture that can support analytics, AI, text processing, data-first application development, and interoperability, as an all-in-one agile cost-savings architecture.
In addition to the successes, Dale comments on the challenges that Health Catalyst has faced under a very ambitious DOS development plan. In its current state, DOS has made some significant improvements to overcome early mistakes, and is now a very solid enterprise data platform. In the interests of industry-wide learning, Sanders will talk transparently about those mistakes and how those learnings are being applied to the DOS platform, positioning it to evolve gracefully over the next 25 years.
View the webinar to learn how the DOS reference architecture:
- Helps manage the 2,000+ compulsory measures in US healthcare
- Enables applications as varied as a real-time patient safety surveillance system, and an activity-based costing system in one platform
- Can ingest data of any type or velocity from over 300 healthcare source systems and growing
- Bundles tools, applications, and analytics that would cost 3-6x more to build on your own
- Compares to EHR vendors as an option to serve as an enterprise data and analytics platform
- Is a performant, sustainable, and maintainable platform for deploying AI models in the natural flow of the healthcare data pipeline
- Provides curated data content and models while still allowing for the agility of a late binding design option
- Functions as a reference architecture that all healthcare organizations and vendors will ultimately have to build in their pursuit of digital health
Knowledge graphs ilaria maresi the hyve 23apr2020Pistoia Alliance
Data for drug discovery and healthcare is often trapped in silos which hampers effective interpretation and reuse. To remedy this, such data needs to be linked both internally and to external sources to make a FAIR data landscape which can power semantic models and knowledge graphs.
How to Use Geospatial Data to Identify CPG Demnd HotspotsCARTO
The combination of new location data streams and spatial data science techniques opens up a new array of opportunities for CPG data and marketing professionals seeking where to prioritize in terms of ramping up distribution and identifying POS (points of sale).
Natural Language Understanding in HealthcareDavid Talby
The ability of software to reason, answer questions and intelligently converse about clinical notes, patient stories or biomedical papers has risen dramatically in the past few years.
This talk covers state of the art natural language processing, deep learning, and machine learning libraries in this space. We'll share benchmarks from industry & research projects on use cases such as clinical data abstraction, patient risk prediction, named entity recognition & resolution, and negation scope detection.
New technologies such as the Internet of Things and Cloud Computing are expected to leverage the current
trend of Smart Farming, introducing more sensors, robots and artificial intelligence, encompassed by the
phenomenon of Big Data.
This presentation will give a quick insight into the state-of-the-art of Big Data applications in Smart Farming
and identify the related challenges that have to be addressed. It shows that the scope of Big Data
applications in Smart Farming goes beyond the farm; it is influencing the entire food supply chain. Big data
are being used to provide predictive insights in farming operations, drive real-time operational decisions, and
redesign business processes for game-changing business models.
It is expected that Big Data will cause major shifts in roles and power relations among different players in
current food supply chain networks. The landscape of stakeholders exhibits an interesting game between
powerful tech companies, venture capitalists and often small startups and new entrants. At the same time
there are several public institutions that publish open data, under the condition that the privacy of persons
must be guaranteed. The future of Smart Farming may unravel in a continuum of two extreme scenarios: 1)
closed, proprietary systems or 2) open, collaborative systems.
The development of data and application infrastructures (platforms and standards) and their institutional
embedment will play a crucial role in the battle between these scenarios. A major challenge is therefore to
cope with governance issues and define suitable business models for data sharing in different supply chain
scenarios.
ICRISAT Global Planning Meeting 2019:Research Program - Innovation Systems fo...ICRISAT
The Global Planning Meeting 2019 focused on an innovation systems approach harnesses the conditions needed to create demand for technologies and creates the knowledge that may be used to bring about such changes…innovations most often emerge from a systems of actors collaborating, communicating and learning, methodologies and tools to create innovations, understand entry points/tradeoffs and leverage actors towards profitable resilient and sustainable agri-food systems at scale and work together to contribute to ICRISAT’s mission.
Unleashing the power of data in transforming livestock agriculture in Ethiopia ILRI
Presented by Fasil Getachew, Setegn Worku, Wondmeneh Esatu and Tadelle Dessie at the 27 Annual Conference of the Ethiopian Society of Animal Production (ESAP), EIAR, Addis Ababa, 29–31 August 2019
Keynote IoT in Agriculture opening academic year CIHEAM ZaragozaSjaak Wolfert
Keynote presentation for the opening of the academic year at CIHEAM institute for Mediterranean agricultural research in Zaragoza. It is about how IoT and Big Data are transforming Agriculture in Europe and what the main challenges are: governance, business models and open infrastructures. This is illustrated from several use cases in the Internet of Food and Farm 2020 (IoF2020) project.
Dutch dairy sector and expectations from open datagodanSec
Frido Hamoen (CRV) presented at the 2nd International Workshop: Creating Impact with Open Data in Agriculture and Nutrition in The Hague, 10 September 2015.
Farm Management System - Delivering a Precision Agriculture SolutionHPCC Systems
Jeff Bradshaw & Graeme McCracken, RBI, present at the 2016 HPCC Systems Engineering Summit Community Day.
In this session, we will share our use case on how we have collected data from remote Farm Management Systems (used by the Farmers/Growers to manage their farms), and overlaying that with weather data and actual machinery data (IoT) and using this data to feed Agronomists and Crop Protection/Seed Manufacturers to get recommendations back. The goal is to deliver a precision agriculture solution which helps the Farmer to increase his yield and helps us to feed the growing population of the world.
Jeff Bradshaw is the founder of Adaptris and Group CTO of Adaptris/F4F/DBT within Reed Business Information. He has spent his career integrating data wherever it resides and in-flight across a number of industries including Agriculture, Airlines, Telecommunications, Healthcare, Government and Finance.
Jeff has worked with and contributed to a number of international standards bodies and continues to work with large enterprises to help them extract value from their data silos and share data seamlessly with their trading partners to achieve business benefit. For the last few years Jeff has been focusing on Big Data and how to gather that across a wide range of sources to help gain insight into the agri-food supply chain.
Graeme is the Chief Operating Officer for Proagrica, the global agricultural and animal health division within RELX covering Media, Software, Integration & Connectivity and Data & Analytics. Prior to this role, Graeme was the CEO of RELX’s Construction Data & Analytics business in North America with a background in data, product and IT innovation across a complex portfolio of companies in Europe, North America and Australasia.
Graeme has been in RELX for 24 years driving a range of strategic initiatives and building strong teams that are well motivated, involved and having fun. As part of overall strategic alignment, successfully delivered the divestment of a number of divisions whilst ensuring that these units were well set for the future. Impressive track record in transforming a range of business units across RELX and setting them on a successful growth path.
Digital innovation for sustainable food systemsSjaak Wolfert
This presentation will show that digital solutions help addressing multiple sustainability issues, particularly illuminating how producers and consumers can use digitalisation to support a transition towards healthier diets.
The sustainable use of animal genetics in developing countriesILRI
Presented by Steve Staal at the 2nd International Conference on Agricultural and Rural Development in Southeast Asia, Manila, Philippines, 12 November 2014
Reflections on making EFSA an open science organisationNikos Manouselis
Slides of talk at the Workshop on e-Infrastructures supporting Food Safety Risk Assessment, hosted by the European Food Safety Authority (EFSA), Parma, May 13th, 2015.
To help reaching the Sustainable Development Goals, CGIAR must tap into Big Data. Within the programme on Climate Change for Agriculture and Food Security (CCAFS), researchers have already applied Big Data analytics to agricultural and weather records in Colombia, revealing how climate variation impacts rice yields. After defining its Open Data-Open Access strategy, CGIAR has launched an internal call for proposals for big data analytics platforms that will provide services to the Agri-Food system programmes and parners, and will interconnect the CGIAR data to other multi-disciplinary big data. The seminar will present the pespectives of the envisioned platforms.
Guidelines on the use of sensors to monitor animal health and productivity; a...Claudia Kamphuis
this presentation was given at the third Sund DairyCare conference in Zadar, Croatie. It discusses the need to have protocols to evaluate sensor technologies for their performance on-farm
De veterinaire dag was onderdeel van het eerste internationale congres over Precision Dairy Farming in het WTC in Leeuwarden. Deze dag was speciaal voor de Nederlands rundvee dierenarts die meer wilde weten over het gebruik van sensoren in de melkveehouderij. Deze presentatie gaat specifiek over het gebruik van uiergezondheidsattenties gegenereerd door melk robots.
Deze presentatie is gehouden tijdens een studiedag over sensoren voor de automatische detective van kreupelheid bij melkvee, georganiseerd door het ILVO in Belgie.
The use of successfull inseminations to avoid the high costs and intensive method of progesterone measurements and to crank up the numbers to evaluate sensitivity of automated heat detection systems
Cows in the cloud, Down to earth, 8-9 September 2015Claudia Kamphuis
Farming involves entrepreneurship, setting milestones and preparing for the future. In addition, farming is continuously subject to change, due to growth, society, regulations, finance, subsidy, etc. Therefore solid advice is key for a sustainable, profitable and enjoyable future in farming. A variety of speakers from different disciplines will share interesting insights and knowledge to help you in supporting farmers to reach their chosen milestones.
activity meters are often used for automated oestrus detection. But is there more benefit from monitoring activity of cows? This presentation was part of the SUND Dairycare conference held in 2015, in Cordoba, Spain
Can we estimate the economic benefit of precision livestock technologiesClaudia Kamphuis
A presentation about a modelling tool to estimate the economic impact of implementing precision livestock technologies (PLF) on farm. Presented at the EAAP/EU-PLF Conference, 2014, in Copenhagen, Denmark
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.
7. Big Data: about V’s and one A
7
A
Volume
VelocityVariety
John Mashley, 1990s
8. Big Data: about V’s and one A
8
A
Volume
Velocity
Variety
VeracityVariability
Value
Visualisation
9. Big Data: about V’s and one A
9
A
Volume
Velocity
Variety
Veracity
Variability
ValueVisualisation
Virality
Viscosity
Validity
...
10. Big Data: about V’s and one A
10
Analytics
Volume
Velocity
Variety
Veracity
Variability
ValueVisualisation
Virality
Viscosity
Validity
...
11. Precision Dairy Farming (PDF)
Sensor or automation technologies to
Reduce human labour
Support (daily) management
Improve farm profitability & sustainability
But also...
Monitor parameters related to health/fertility of individual cows
Automatic detection of events (e.g. estrus and mastitis)
11
12. Big Data = PDF?
12
Analytics
Volume
Velocity
Variety
VeracityVariability
Value
Visualisation
13. Big Data = PDF?
13
Analytics
Volume
Velocity
Variety
VeracityVariability
Value
Visualisation
14. Big Data = PDF?
14
Analytics
Volume
Velocity
Variety
VeracityVariability
Value
Visualisation
15. Big Data = PDF?
15
Analytics
Volume
Velocity
Variety
Veracity
Variability
Visualise
Value
16. Big Data ≠ PDF?
16
Analytics
Volume
Velocity
Variety
Veracity
Variability
Visualise
Value
17. Data many cows/farms for many
years
Sources data reside and originate
outside farm fence
Interaction across farms and
through chain
Complex interaction across
several factors
Aggregate data from many
sources
Data few cows (herd)/ 1 farm
.......................................
Sources of data reside and
originate on cow / on-farm........
Tools for farm management
.........
Research is focused on one factor
.....
Data from a few sources, or from
same tech provider .................
Big Data ≠ PDF
17
PDFBig Data
18. Big (Dairy) Data projects @WUR
18
Estimating cow individual feed intake
Smart animal breeding
Predict longevity
Gentore
Tools for resilience & efficiency
19. Estimating cow individual feed intake
More efficient food production required to feed the world
More efficient cows needed
Improve feed efficiency
20. Estimating cow individual feed intake
More efficient food production required to feed the world
More efficient cows needed
Improve feed efficiency
22. Big Data components
15 years data
7 farms
1850 cows
Variety of sources
Machine learning
22
Analytics
Volume
Velocity
Variety
Veracity
Variability
Visualise
Value
23. STW-Breed4Food
Partnership Programme
Smart animal breeding: predict longevity
Survival to second lactation + one week
complex/summary trait from several factors
economically important
early prediction of survival allows for improved
breeding and culling management decisions
23
24. Big Data components
24
Machine learning
Several sources
genomic, phenotypic,
environmental,
sensor
Large number of
animals
Analytics
Volume
Velocity
Variety
VeracityVariability
Visualise
Value
25. Horizon 2020
Gentore:
tools for resilience & efficiency
Develop innovative tools to optimise resilience and
efficiency in widely varying environments
beef, milk, and mixed environments
management tools for on-farm assessment for animal resilience
to allow improved breeding and culling decisions
Develop breeding strategies
Provide policy support by using tools to
compare future incentive/risk scenarios
25
26. Big Data components
Multidisciplinary
scientific teams
Chain level
- breeding organisations
- farm tech companies
- advisory services (e.g vet)
- farm sectors
Data base including
>1 million genotypes
26
Analytics
Volume
Velocity
Variety
Veracity
Variability
Visualise
Value
27. Take home messages around Big Data
Big Data is more than just a lot of data (volume)
Big Data is a rookie within livestock sector
Not everybody is doing it yet
PDF is a data source for Big Data
Added value of Big Data is yet to be proven
27
So, Big Data.
If you have not heard about this by now, you must have been on a long holiday without wifi.
You all must be familiar with the famous stories of facebook, google and twitter all collecting and analysing huge volumes of data, turning it into information and into profit.
Or the use of big data by retail, keeping records on things people purchase and use this information for logistics, or to decide which products to discount, and even where to put products in the store.
Banks use bank account data to create products that better suit their clients
UPS is making reference to using Big Data approach to optimize their routing which saved them millions on fuel costs on an annual basis
All in all, Big Data is known via these large companies, and all state that BigData made them rich.
Big Data stands for combining large amounts of data in smart ways to improve or optimise management and success is guaranteed.
And because everybody seems to be working with big data, it’s definately a buzzword.
So, Big Data.
If you have not heard about this by now, you must have been on a long holiday without wifi.
You all must be familiar with the famous stories of facebook, google and twitter all collecting and analysing huge volumes of data, turning it into information and into profit.
Or the use of big data by retail, keeping records on things people purchase and use this information for logistics, or to decide which products to discount, and even where to put products in the store.
Banks use bank account data to create products that better suit their clients
UPS is making reference to using Big Data approach to optimize their routing which saved them millions on fuel costs on an annual basis
All in all, Big Data is known via these large companies, and all state that BigData made them rich.
Big Data stands for combining large amounts of data in smart ways to improve or optimise management and success is guaranteed.
And because everybody seems to be working with big data, it’s definately a buzzword.
When you place Big Data in an agricultural setting, I think that this picture summarizes Big Data.
Or I should be more precise, the driver of this tractor symbolises the concept of Big Data.Why?
Well, the driver is invisible, similar as Big Data is invisible, and thus both are a bit mysterious...
More importantly, I think that what the driver is doing, is similar to what Big Data is doing. The driver combines information from several sources, all having different formats, some come in continuously, others less frequent but the driver is using all this information near real-time to adjust his route, or the amount of fertilizer, or whatever he’s doing on this tractor. He’s adjusting or is making decisions based on the realtime information coming in.
As in many sectors, Big Data is of increasing interest livestock, because also here the amount of data that is being generated is continuously growing, partly because of the increasing amounts of sensor technologies. We have sensors monitoring how active a cow is, where she is, how much she’s eating, how much milk she’s producing and with what composition, we have the information around the DNA of cows. We even have information of parcels, how much grass is growing, how are they fertilized etc etc....all this increasing volumes of data are food for thoughts, and makes Big Data
We also have increasing amounts of data of parcels, how much product they are generating via drone images.
When you place Big Data in an agricultural setting, I think that this picture summarizes Big Data.
Or I should be more precise, the driver of this tractor symbolises the concept of Big Data.Why?
Well, the driver is invisible, similar as Big Data is invisible, and thus both are a bit mysterious...
More importantly, I think that what the driver is doing, is similar to what Big Data is doing. The driver combines information from several sources, all having different formats, some come in continuously, others less frequent but the driver is using all this information near real-time to adjust his route, or the amount of fertilizer, or whatever he’s doing on this tractor. He’s adjusting or is making decisions based on the realtime information coming in.
As in many sectors, Big Data is of increasing interest livestock, because also here the amount of data that is being generated is continuously growing, partly because of the increasing amounts of sensor technologies. We have sensors monitoring how active a cow is, where she is, how much she’s eating, how much milk she’s producing and with what composition, we have the information around the DNA of cows. We even have information of parcels, how much grass is growing, how are they fertilized etc etc....all this increasing volumes of data are food for thoughts, and makes Big Data
We also have increasing amounts of data of parcels, how much product they are generating via drone images.
But, how is Big Data defined. What is it exactly?
The tern Big Data is probably introduced by John Mashely in the early 1990s, where he used the term to warn the high-tech community of that time that challenges were about to occur with the continuous advances in computer storage, processing power, and developments in the generation of data. It took some years before the three characteristics got associated with it to form today’s mainstream definition.
These characteristics are the first, and most commonly known, three Vs.
Firstly, Volume. I deliberately increased the size of this circle because Volume is likely the first thing people think about when talking about Big Data. It’s BIG. It’s the characteristic that is often emphasized in the media, and it’s also the characteristic to impress people the easiest. At the same time, the threshold of what the volume should be to be called Big is unknown, and the threshold is subjective and dependent on the industry and application. It may be anything whose size exceeds the ability of typical software used to capture, store, manage, and analyse or any attribute challenging the constraints of a system capability or business needs. Others call it Big when data sets are so large or complex that traditional data processing applications are inadequate. And to add a bit of complexity, data we consider Big today may not be considered Big tomorrow due to the continuous advances in processing, storage, and other system capabilities.
The other two characteristics are
Velocity referring to the Data component of Big Data. It links to how frequently data are generated, the increasing speed of data arrival and processing, and the ability to respond to events as they occur
and
Variety also referring to the Data component of Big Data. It links to the variety of incompatible formats, non-aligned structures, and inconsistent semantics and this variety in data sources have been mentioned as the greatest barrier to effective data analysis.
Over the years more Vs were added to define Big Data.
Veracity: refering to the quality of data, the need to trust the data
Variability: variance in meaning in sentiment analysis or the inconsistency of data
Visualization: creation of complex graphs
Value: the value of relying on data-driven analysis for decision - making
And even more without being explained at all.....and of course we can expect some new ones in the future too.
So we have a lot of Vs.....and we still have one A
That A stands for analytics.
Collecting more data that meet one or more of the Vs mentioned is still worthless if you don’t do anything with it. Analytics are required to generate the insights that enable improved decision-making.
So, that’s Big Data. And I just want to stress at this point that although we do have many Vs that define BD, none of these Vs have a threshold nor is there a threshold to the number of Vs that needs to be fulfilled to call something BD.....
Then.....what is Precision dairy farming?
Well, we have heard a lot today about this topic, so this slide is just a brief summary of what PDF is.
I have been working in this field of PDF the past years, and when I just started to hear from Big Data, and saw the enthusiasm that this word caused by many people, I couldn’t really understand the fuzz around it.
I figured that my research within the field of Precision Dairy Farming was basically Big Data.
I mean, when working with PDF technologies, you work with large amounts of data generated by sensors
You are already working with data that are generated at high speed, true.
For example, with automatic milking, electrical conductivity can be measured and recorded every second.....
But....the need for realtime processing and descision making in PDF area so far has not been that much of an issue....
Based on sensor measurements in robotic milkings, the decision to automatically draft milk with blood is probably the only example where a decision has to be taken immediately (after milking).
All other events monitored by technologies do not require instant actions.
And with the PDF technologies you have to face the challenges with differences in veracity? Activity from one sensor technology is different from another technology and in you’re analysis, and in the interpretation of it, you have to know these differences and the consequences of it.
And by using PDF technologies, we often have tools that monitor stuff, so yup, we can tick that box, or circle too, perhaps with a somewhat lessen extend....
All the other Vs that define Big Data, are not yet used by PDF, or to a very basic or small extend.
So by thinking about this....and by looking at this picture I came to realize that PDF is not Big Data.
It certainly has key elements from BD, and new analytical tools like machine learning or data mining are required to understand and visualize the data, but they are different
And although I just mentioned that there is no threshold set at the number of Vs that need to be fulfilled before you can call something BD I wasn’t so much impressed by
This picture myself....
So apparently, I have not been doing as much of Big Data than I thought before...
Thinking about this a bit longer, I would like to highlight some more differences between BD and PDF.
Now, why is this important to recognize? Because I belief that is you think you’re doing something that is actually not Big Data, you will stick with the same procedures and methodologies to collect preprocess and analyse data, you stick with the same approach to define research questions and by doing so you will not make the progress that Big Data is actually offering.
But, as mentioned earlier....Big Data is buzzing around, and developments in sensor technologies, or precision dairy farming, are pushing these developments of Big data in the livestock sector too.
I’d like to explain three projects we are currently working on at WUR that are related to dairy, and that all work in the Big Data field.
For each of these projects, I briefly explain the goal, and I would like to demonstrate why these projects are more than just PDF
Plaatje onderzoeksvoorstel
1) multidisciplinary scientific expertise in genomics, environmental assessment, nutritional physiology, health management, precision livestock farming, mathematical modelling, and socio-economics; 2) partners and stakeholders representing breeding organisations, farm technology companies, farm and veterinary advisory services, and farm sectors (organic, grazing, etc.); and 3) a unique data basis including >1 million genotypes