In this deck from the HPC User Forum in Milwaukee, Tim Barr from Cray presents: Perspective on HPC-enabled AI.
"Cray’s unique history in supercomputing and analytics has given us front-line experience in pushing the limits of CPU and GPU integration, network scale, tuning for analytics, and optimizing for both model and data parallelization. Particularly important to machine learning is our holistic approach to parallelism and performance, which includes extremely scalable compute, storage and analytics."
Watch the video: https://wp.me/p3RLHQ-hpw
Learn more: http://cray.com
Sign up for our insideHPC Newsletter: http://insidehpc.com/newsletter
Digital Transformation - #StrataData London 2017 - Data101Ellen Friedman
Presented at Strata Data London conference May 2017 in the Data 101 track, this presentation explores what is needed in planning, architecture, and cultural organization for effective digital transformation.
For the full video of this presentation, please visit:
http://www.embedded-vision.com/platinum-members/embedded-vision-alliance/embedded-vision-training/videos/pages/may-2015-embedded-vision-summit-baidu
For more information about embedded vision, please visit:
http://www.embedded-vision.com
Dr. Ren Wu, former distinguished scientist at Baidu's Institute of Deep Learning (IDL), presents the keynote talk, "Enabling Ubiquitous Visual Intelligence Through Deep Learning," at the May 2015 Embedded Vision Summit.
Deep learning techniques have been making headlines lately in computer vision research. Using techniques inspired by the human brain, deep learning employs massive replication of simple algorithms which learn to distinguish objects through training on vast numbers of examples. Neural networks trained in this way are gaining the ability to recognize objects as accurately as humans.
Some experts believe that deep learning will transform the field of vision, enabling the widespread deployment of visual intelligence in many types of systems and applications. But there are many practical problems to be solved before this goal can be reached. For example, how can we create the massive sets of real-world images required to train neural networks? And given their massive computational requirements, how can we deploy neural networks into applications like mobile and wearable devices with tight cost and power consumption constraints?
In this talk, Ren shares an insider’s perspective on these and other critical questions related to the practical use of neural networks for vision, based on the pioneering work being conducted by his former team at Baidu.
Note 1: Regarding the ImageNet results included in this presentation, the organizers of the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) have said: “Because of the violation of the regulations of the test server, these results may not be directly comparable to results obtained and reported by other teams.” (http://www.image-net.org/challenges/LSVRC/announcement-June-2-2015)
Note 2: The presenter, Ren Wu, has told the Embedded Vision Alliance that “There was some ambiguity with the rules. According to the ‘official’ interpretation of the rules, there should be no more than 52 submissions within a half year. For us, we achieved the reported results after 200 tests total within a half year. We believe there is no way to obtain any measurable gains, nor did we try to obtain any gains, from an 'extra' hundred tests as our networks have billions of parameters and are trained by tens of billions of training samples.”
In this deck from the HPC User Forum in Milwaukee, Tim Barr from Cray presents: Perspective on HPC-enabled AI.
"Cray’s unique history in supercomputing and analytics has given us front-line experience in pushing the limits of CPU and GPU integration, network scale, tuning for analytics, and optimizing for both model and data parallelization. Particularly important to machine learning is our holistic approach to parallelism and performance, which includes extremely scalable compute, storage and analytics."
Watch the video: https://wp.me/p3RLHQ-hpw
Learn more: http://cray.com
Sign up for our insideHPC Newsletter: http://insidehpc.com/newsletter
Digital Transformation - #StrataData London 2017 - Data101Ellen Friedman
Presented at Strata Data London conference May 2017 in the Data 101 track, this presentation explores what is needed in planning, architecture, and cultural organization for effective digital transformation.
For the full video of this presentation, please visit:
http://www.embedded-vision.com/platinum-members/embedded-vision-alliance/embedded-vision-training/videos/pages/may-2015-embedded-vision-summit-baidu
For more information about embedded vision, please visit:
http://www.embedded-vision.com
Dr. Ren Wu, former distinguished scientist at Baidu's Institute of Deep Learning (IDL), presents the keynote talk, "Enabling Ubiquitous Visual Intelligence Through Deep Learning," at the May 2015 Embedded Vision Summit.
Deep learning techniques have been making headlines lately in computer vision research. Using techniques inspired by the human brain, deep learning employs massive replication of simple algorithms which learn to distinguish objects through training on vast numbers of examples. Neural networks trained in this way are gaining the ability to recognize objects as accurately as humans.
Some experts believe that deep learning will transform the field of vision, enabling the widespread deployment of visual intelligence in many types of systems and applications. But there are many practical problems to be solved before this goal can be reached. For example, how can we create the massive sets of real-world images required to train neural networks? And given their massive computational requirements, how can we deploy neural networks into applications like mobile and wearable devices with tight cost and power consumption constraints?
In this talk, Ren shares an insider’s perspective on these and other critical questions related to the practical use of neural networks for vision, based on the pioneering work being conducted by his former team at Baidu.
Note 1: Regarding the ImageNet results included in this presentation, the organizers of the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) have said: “Because of the violation of the regulations of the test server, these results may not be directly comparable to results obtained and reported by other teams.” (http://www.image-net.org/challenges/LSVRC/announcement-June-2-2015)
Note 2: The presenter, Ren Wu, has told the Embedded Vision Alliance that “There was some ambiguity with the rules. According to the ‘official’ interpretation of the rules, there should be no more than 52 submissions within a half year. For us, we achieved the reported results after 200 tests total within a half year. We believe there is no way to obtain any measurable gains, nor did we try to obtain any gains, from an 'extra' hundred tests as our networks have billions of parameters and are trained by tens of billions of training samples.”
Q: Can I simply hire one rockstar data scientist to cover all this kind of work?
A: No, interdisciplinary work requires teams
A: Hire leads who can speak the lingo of each required discipline
A: Hire individual contributors who cover 2+ roles, when possible
Statistical Thinking – Solve the Whole Problem
BONUS: Meta Organization – Integration with Adjacent Teams
Co-authors Allen Day @allenday and Paco Nathan @pacoid
Presentation that I delivered at "Accelerate AI, Europe 2018" in London on Sept 19, 2018. My focus is on socio-cultural perspective as well as proving information about various tools, vendors and partners available to help companies get started using AI.
Cloud computing & big data for service innovation & learning2016
Cloud Computing and Big Data for Service Innovations & Learning
Up till now, most of the adoption of cloud computing focusses on the automation and consolidation of traditional IT services. As such, the gains are confined to the uniformity of control, cost reduction and better governance. Recent adoption of the cloud has gradually moved into tactical and even strategic levels thereby demonstrating a high level of gains for using the cloud for business transformations and innovations. Such benefits include dynamism in business model compositions and speed and ease in orchestrating service innovations in the cloud. This talk will shed light on how massive and rapid accumulation of data in the cloud can support human-machine cooperative problem solving and re-define the landscape of Open Innovation and Connectionist Learning via a Knowledge Cloud.
The Proliferation of New Database Technologies and Implications for Data Scie...Domino Data Lab
In this talk, we’ll describe NoSQL (“not-only SQL”) and document-oriented databases and the value they provide for data science companies like Uptake. We will walk through the unique challenges such datastores pose for data science workflows. To make these challenges and lessons learned concrete, we’ll explore data science workflows through a discussion of the development efforts that led to “uptasticsearch”, an R package released by the Uptake Data Science team to reduce friction in interacting with a document store called Elasticsearch. The talk will conclude with a discussion of recent developments in NoSQL technologies and implications for data scientists.
Adoption of the R language has grown rapidly in the last few years, and is ranked as the number-one data science language in several surveys. This accelerating R adoption curve has been driven by the Big Data revolution, and the fact that so many data scientists — having learned R at university — are actively unlocking the secrets hidden in these new, vast data troves. In more than 6 years of writing for the Revolutions blog, I’ve discovered hundreds of applications of R in business, in government, and in the non-profit sector. Sometimes the use of R is obvious, and sometimes it takes a little bit of detective work to learn how R is operating behind the scenes. In this talk, I'll recount some of my favourite applications of R, and show how R is behind some amazing innovations in today’s world.
7 Habits for Big Data in Production - keynote Big Data London Nov 2018Ellen Friedman
You can improve your chances for success with data intensive large scale applications (AI, machine learning and analytics) in production.
This keynote presentation from Big Data London shows you how.
Data is both our most valuable asset and our biggest ongoing challenge. As data grows in volume, variety and complexity, across applications, clouds and siloed systems, traditional ways of working with data no longer work.
Unlike traditional databases, which arrange data in rows, columns and tables, Neo4j has a flexible structure defined by stored relationships between data records.
We'll discuss the primary use cases for graph databases
Explore the properties of Neo4j that make those use cases possible
Look into the visualisation of graphs
Introduce how to write queries.
Webinar, 23 July 2020
Vertical is the New Horizontal - MinneAnalytics 2016 Sri Ambati Keynote on AISri Ambati
Data is the only vertical, Machine Learning, bigdata, artificial intelligence
- Powered by the open source machine learning software H2O.ai. Contributors welcome at: https://github.com/h2oai
- To view videos on H2O open source machine learning software, go to: https://www.youtube.com/user/0xdata
Big Data Overview for Chinese University of Hong Kong Centre for Innovation a...orcsab
This is the student-focused version of the talk I have been delivering in Hong Kong lately. Very similar to the talk I gave to the CUHK Statistics department. In this one I provide a general description of big data, talk about enabling technologies, give examples of big data in action, and talk about big and open data projects right here in Hong Kong.
For more info: http://infoincog.com
This talk shows practical methods for find changes in a variety of kinds of data as well as giving real-world examples from finance, telecom, systems monitoring and natural language processing.
ML Workshop 2: Machine Learning Model Comparison & EvaluationMapR Technologies
How Rendezvous Architecture Improves Evaluation in the Real World
In this addition of our machine learning logistics webinar series we build on the ideas of the key requirements for effective management of machine learning logistics presented in the Overview webinar and in Part I Workshop. Here we focus on model-to-model comparison & evaluation, use of decoy models and more. Listen here: http://info.mapr.com/machine-learning-workshop2.html?_ga=2.35695522.324200644.1511891424-416597139.1465233415
Q: Can I simply hire one rockstar data scientist to cover all this kind of work?
A: No, interdisciplinary work requires teams
A: Hire leads who can speak the lingo of each required discipline
A: Hire individual contributors who cover 2+ roles, when possible
Statistical Thinking – Solve the Whole Problem
BONUS: Meta Organization – Integration with Adjacent Teams
Co-authors Allen Day @allenday and Paco Nathan @pacoid
Presentation that I delivered at "Accelerate AI, Europe 2018" in London on Sept 19, 2018. My focus is on socio-cultural perspective as well as proving information about various tools, vendors and partners available to help companies get started using AI.
Cloud computing & big data for service innovation & learning2016
Cloud Computing and Big Data for Service Innovations & Learning
Up till now, most of the adoption of cloud computing focusses on the automation and consolidation of traditional IT services. As such, the gains are confined to the uniformity of control, cost reduction and better governance. Recent adoption of the cloud has gradually moved into tactical and even strategic levels thereby demonstrating a high level of gains for using the cloud for business transformations and innovations. Such benefits include dynamism in business model compositions and speed and ease in orchestrating service innovations in the cloud. This talk will shed light on how massive and rapid accumulation of data in the cloud can support human-machine cooperative problem solving and re-define the landscape of Open Innovation and Connectionist Learning via a Knowledge Cloud.
The Proliferation of New Database Technologies and Implications for Data Scie...Domino Data Lab
In this talk, we’ll describe NoSQL (“not-only SQL”) and document-oriented databases and the value they provide for data science companies like Uptake. We will walk through the unique challenges such datastores pose for data science workflows. To make these challenges and lessons learned concrete, we’ll explore data science workflows through a discussion of the development efforts that led to “uptasticsearch”, an R package released by the Uptake Data Science team to reduce friction in interacting with a document store called Elasticsearch. The talk will conclude with a discussion of recent developments in NoSQL technologies and implications for data scientists.
Adoption of the R language has grown rapidly in the last few years, and is ranked as the number-one data science language in several surveys. This accelerating R adoption curve has been driven by the Big Data revolution, and the fact that so many data scientists — having learned R at university — are actively unlocking the secrets hidden in these new, vast data troves. In more than 6 years of writing for the Revolutions blog, I’ve discovered hundreds of applications of R in business, in government, and in the non-profit sector. Sometimes the use of R is obvious, and sometimes it takes a little bit of detective work to learn how R is operating behind the scenes. In this talk, I'll recount some of my favourite applications of R, and show how R is behind some amazing innovations in today’s world.
7 Habits for Big Data in Production - keynote Big Data London Nov 2018Ellen Friedman
You can improve your chances for success with data intensive large scale applications (AI, machine learning and analytics) in production.
This keynote presentation from Big Data London shows you how.
Data is both our most valuable asset and our biggest ongoing challenge. As data grows in volume, variety and complexity, across applications, clouds and siloed systems, traditional ways of working with data no longer work.
Unlike traditional databases, which arrange data in rows, columns and tables, Neo4j has a flexible structure defined by stored relationships between data records.
We'll discuss the primary use cases for graph databases
Explore the properties of Neo4j that make those use cases possible
Look into the visualisation of graphs
Introduce how to write queries.
Webinar, 23 July 2020
Vertical is the New Horizontal - MinneAnalytics 2016 Sri Ambati Keynote on AISri Ambati
Data is the only vertical, Machine Learning, bigdata, artificial intelligence
- Powered by the open source machine learning software H2O.ai. Contributors welcome at: https://github.com/h2oai
- To view videos on H2O open source machine learning software, go to: https://www.youtube.com/user/0xdata
Big Data Overview for Chinese University of Hong Kong Centre for Innovation a...orcsab
This is the student-focused version of the talk I have been delivering in Hong Kong lately. Very similar to the talk I gave to the CUHK Statistics department. In this one I provide a general description of big data, talk about enabling technologies, give examples of big data in action, and talk about big and open data projects right here in Hong Kong.
For more info: http://infoincog.com
This talk shows practical methods for find changes in a variety of kinds of data as well as giving real-world examples from finance, telecom, systems monitoring and natural language processing.
ML Workshop 2: Machine Learning Model Comparison & EvaluationMapR Technologies
How Rendezvous Architecture Improves Evaluation in the Real World
In this addition of our machine learning logistics webinar series we build on the ideas of the key requirements for effective management of machine learning logistics presented in the Overview webinar and in Part I Workshop. Here we focus on model-to-model comparison & evaluation, use of decoy models and more. Listen here: http://info.mapr.com/machine-learning-workshop2.html?_ga=2.35695522.324200644.1511891424-416597139.1465233415
The logistics of machine learning typically take waaay more effort than the machine learning itself. Moreover, machine learning systems aren't like normal software projects so continuous integration takes on new meaning.
The Hive Think Tank: Rendezvous Architecture Makes Machine Learning Logistics...The Hive
Think Tank Event 10/23/2017, hosted by The Hive and presented by Ted Dunning, Chief Application Architect of MapR Technologies and Ellen Friedman of MapR Technologies.
You know that a single number isn't a good summary of a measurement. T-digest and other non-uniform histograms can make it easy to keep track of an entire distribution and can be combined in OLAP queries.
The latest t-digest is faster, more accurate and has hard bounds on size.
Machine Learning Success: The Key to Easier Model ManagementMapR Technologies
Join Ellen Friedman, co-author (with Ted Dunning) of a new short O’Reilly book Machine Learning Logistics: Model Management in the Real World, to look at what you can do to have effective model management, including the role of stream-first architecture, containers, a microservices approach and a DataOps style of work. Ellen will provide a basic explanation of a new architecture that not only leverages stream transport but also makes use of canary models and decoy models for accurate model evaluation and for efficient and rapid deployment of new models in production.
This talk focuses on how larger data sets are not only enabling advanced techniques, but also increasing the number of problems within reach of relatively simple techniques, that is "cheap learning".
Complement Deep Learning with Cheap Learning: Recent results of deep learning on hard problems has set the data world all a titter and made deep learning the fashion of the time.
But it is very important to remember that as data expands, the learning problems that are encountered are often nearly green field problems and it is often possible to solve these problems using remarkably simple techniques. Indeed, on many problems these simple techniques will give results as good as more complex ones, not because they are profound, but because many problems become simpler at scale.
That said, it isn’t always obvious how to do this. I will describe some of these techniques and show how they can be applied in practice.
ML Workshop 1: A New Architecture for Machine Learning LogisticsMapR Technologies
Having heard the high-level rationale for the rendezvous architecture in the introduction to this series, we will now dig in deeper to talk about how and why the pieces fit together. In terms of components, we will cover why streams work, why they need to be persistent, performant and pervasive in a microservices design and how they provide isolation between components. From there, we will talk about some of the details of the implementation of a rendezvous architecture including discussion of when the architecture is applicable, key components of message content and how failures and upgrades are handled. We will touch on the monitoring requirements for a rendezvous system but will save the analysis of the recorded data for later. Listen to the webinar on demand: https://mapr.com/resources/webinars/machine-learning-workshop-1/
State of the Art Robot Predictive Maintenance with Real-time Sensor DataMathieu Dumoulin
Our Strata Beijing 2017 presentation slides where we show how to use data from a movement sensor, in real-time, to do anomaly detection at scale using standard enterprise big data software.
Tensor Abuse - how to reuse machine learning frameworksTed Dunning
Tensors are a very useful tool for mathematical programming. Moreover, the optimization frameworks that are part of most machine learning frameworks have some very cool uses outside of the normal machine learning kinds of tasks.
Ellen Friedman and I spoke at the ACM meetup about how stream-first architecture can have a big impact and how the logistics of machine learning is a great example of that impact.
This is my half of the presentation.
Fighting financial fraud at Danske Bank with artificial intelligenceRon Bodkin
Danske Bank, the leader in mobile payments in Denmark, is innovating with AI. Danske Bank’s existing fraud detection engine is being enhanced with deep learning algorithms that can analyze potentially tens of thousands of latent features. Danske Bank’s current system is largely based on handcrafted rules created by the business, based on intuition and some light analysis. The system is effective at blocking fraud, but it has a high rate of false positives, which is expensive and inconvenient, and it has proved impractical to update and maintain as fraudsters evolve their capabilities. Moreover, the bank understands that fraud is getting worse in the near- and long-term future due to the increased digitization of banking and the prevalence of mobile banking applications and recognizes the need to use cutting-edge techniques to engage fraudsters not where they are today but where they will be tomorrow.
Application fraud is an important emerging trend, in which machines fill in transaction forms. There is evidence that criminals are employing sophisticated machine-learning techniques to attack, so it’s critical to use sophisticated machine learning to catch fraud in banking and mobile payment transactions.
Ron Bodkin and Nadeem Gulzar explore how Danske Bank uses deep learning for better fraud detection. Danske Bank’s multistep program first productionizes “classic” machine learning techniques (boosted decision trees) while in parallel developing deep learning models with TensorFlow as a “challenger” to test. The system was first tested in shadow production and then in full production in a champion-challenger setup against live transactions. Ron and Nadeem explain how the bank is integrating the models with the efforts already running, giving the bank and its investigation team the ability to adapt to new patterns faster than before and taking on complex highly varying functions not present in the training examples.
Ted Dunning, Chief Application Architect, MapR at MLconf SFMLconf
Abstract: Near real-time Updates for Cooccurrence-based Recommenders
Most recommendation algorithms are inherently batch oriented and require all relevant history to be processed. In some contexts such as music, this does not cause significant problems because waiting a day or three before recommendations are available for new items doesn’t significantly change their impact. In other contexts, the value of items drops precipitously with time so that recommending day-old items has little value to users.
In this talk, I will describe how a large-scale multi-modal cooccurrence recommender can be extended to include near real-time updates. In addition, I will show how these real-time updates are compatible with delivery of recommendations via search engines.
Similar to Big Data LDN 2017: Machine Learning: What Works And What They Won’t Tell You (20)
Blueprint Series: Banking In The Cloud – Ultra-high Reliability ArchitecturesMatt Stubbs
Data architecture for a challenger bank.Speaker: Jason Maude, Head of Technology Advocacy, Starling BankSpeaker Bio: Jason Maude is a coder, coach, and public speaker. He has over a decade of experience working in the financial sector, primarily in creating and delivering software. He is passionate about explaining complex technical concepts to those who are convinced that they won't be able to understand them. He currently works at Starling Bank as their Head of Technology Advocacy and host of the Starling podcast.Filmed at Skills Matter/Code Node London on 9th May 2019 as part of the Big Data LDN Meetup Blueprint Series.Meetup sponsored by DataStax.
Speed Up Your Apache Cassandra™ Applications: A Practical Guide to Reactive P...Matt Stubbs
Speaker: Cedrick Lunven, Developer Advocate, DataStax
Speaker Bio: Cedrick is a Developer Advocate at DataStax where he finds opportunities to share his passions by speaking about developing distributed architectures and implementing reference applications for developers. In 2013, he created FF4j, an open source framework for Feature Toggle which he still actively maintains. He is now contributor in JHipster team.
Talk Synopsis: We have all introduced more or less functional programming and asynchronous operations into our applications in order to speed up and distribute treatments (e.g., multi-threading, future, completableFuture, etc.). To build truly non-blocking components, optimize resource usage, and avoid "callback hell" you have to think reactive—everything is an event.
From the frontend UI to database communications, it’s now possible to develop Java applications as fully reactive with frameworks like Spring WebFlux and Reactor. With high throughput and tunable consistency, applications built on top of Apache Cassandra™ fit perfectly within this pattern.
DataStax has been developing Apache Cassandra drivers for years, and in the latest version of the enterprise driver we introduced reactive programming.
During this session we will migrate, step by step, a vanilla CRUD Java service (SpringBoot / SpringMVC) into reactive with both code review and live coding. Bring home a working project!
Filmed at Skills Matter/Code Node London on 9th May 2019 as part of the Big Data LDN Meetup Blueprint Series.
Meetup sponsored by DataStax.
Blueprint Series: Expedia Partner Solutions, Data PlatformMatt Stubbs
Join Anselmo for an engaging overview of the new end-to-end data architecture at Expedia Group, taking a journey through cloud and on-prem data lakes, real-time and batch processes and streamlined access for data producers and consumers. Find out how the new architecture unifies a complex mix of data sources and feeds the data science development cycle. Expedia might appear to be a market-leading travel company – in reality, it’s a highly successful technology and data science company.
Blueprint Series: Architecture Patterns for Implementing Serverless Microserv...Matt Stubbs
Richard Freeman talks about how the data science team at JustGiving built KOALA, a fully serverless stack for real-time web analytics capture, stream processing, metrics API, and storage service, supporting live data at scale from over 26M users. He discusses recent advances in serverless computing, and how you can implement traditionally container-based microservice patterns using serverless-based architectures instead. Deploying Serverless in your organisation can dramatically increase the delivery speed, productivity and flexibility of the development team, while reducing the overall running, DevOps and maintenance costs.
Big Data LDN 2018: DATABASE FOR THE INSTANT EXPERIENCEMatt Stubbs
Date: 14th November 2018
Location: Customer Experience Theatre
Time: 12:30 - 13:00
Speaker: David Maitland
Organisation: Redis Labs
About: This session will cover the technology underpinning at the software infrastructure level required to deliver the instant experience to the end user and enterprises alike. Use cases and value derived by major brands will be shared in this insightful session based the world's most loved database REDIS.
Big Data LDN 2018: BIG DATA TOO SLOW? SPRINKLE IN SOME NOSQLMatt Stubbs
Date: 14th November 2018
Location: Customer Experience Theatre
Time: 11:50 - 12:20
Speaker: Perry Krug
Organisation: Couchbase
About: Who wants to see an ad today for the shoes they bought last week? Everyone knows that customer experience is driven by data: don't waste an opportunity to get them the right data at the right time. Real-time results are critical, but raw speed isn't everything: you need power and flexibility to react to changes on the fly. Come learn how market-leading enterprises are using Couchbase as their speed layer for ingestion, incremental view and presentation layers alongside Kafka, Spark and Hadoop to liberate their data lakes.
Big Data LDN 2018: ENABLING DATA-DRIVEN DECISIONS WITH AUTOMATED INSIGHTSMatt Stubbs
Date: 13th November 2018
Location: Customer Experience Theatre
Time: 11:50 - 12:20
Speaker: Charlotte Emms
Organisation: seenit
About: How do you get your colleagues interested in the power of data? Taking you through Seenit’s journey using Couchbase's NoSQL database to create a regular, fully automated update in an easily digestible format.
Big Data LDN 2018: DATA MANAGEMENT AUTOMATION AND THE INFORMATION SUPPLY CHAI...Matt Stubbs
Date: 14th November 2018
Location: Governance and MDM Theatre
Time: 10:30 - 11:00
Speaker: Mike Ferguson
Organisation: IBS
About: For most organisations today, data complexity has increased rapidly. In the area of operations, we now have cloud and on-premises OLTP systems with customers, partners and suppliers accessing these applications via APIs and mobile apps. In the area of analytics, we now have data warehouse, data marts, big data Hadoop systems, NoSQL databases, streaming data platforms, cloud storage, cloud data warehouses, and IoT-generated data being created at the edge. Also, the number of data sources is exploding as companies ingest more and more external data such as weather and open government data. Silos have also appeared everywhere as business users are buying in self-service data preparation tools without consideration for how these tools integrate with what IT is using to integrate data. Yet new regulations are demanding that we do a better job of governing data, and business executives are demanding more agility to remain competitive in a digital economy. So how can companies remain agile, reduce cost and reduce the time-to-value when data complexity is on the up?
In this session, Mike will discuss how companies can create an information supply chain to manufacture business-ready data and analytics to reduce time to value and improve agility while also getting data under control.
Date: 13th November 2018
Location: Governance and MDM Theatre
Time: 12:30 - 13:00
Organisation: Immuta
About: Artificial intelligence is rising in importance, but it’s also increasingly at loggerheads with data protection regimes like the GDPR—or so it seems. In this talk, Sophie will explain where and how AI and GDPR conflict with one another, and how to resolve these tensions.
Big Data LDN 2018: REALISING THE PROMISE OF SELF-SERVICE ANALYTICS WITH DATA ...Matt Stubbs
Date: 13th November 2018
Location: Governance and MDM Theatre
Time: 11:50 - 12:20
Speaker: Mark Pritchard
Organisation: Denodo
About: Self-service analytics promises to liberate business users to perform analytics without the assistance of IT, and this in turn promises to free IT to focus on enhancing the infrastructure.
Join us to learn how data virtualization will allow you to gain real-time access to enterprise-wide data and deliver self-service analytics. We will explore how you can seamlessly unify fragmented data, replace your high-maintenance and high cost data integrations with a single, low-maintenance data virtualization layer; and how you can preserve your data integrity and ensure data lineage is fully traceable.
Big Data LDN 2018: TURNING MULTIPLE DATA LAKES INTO A UNIFIED ANALYTIC DATA L...Matt Stubbs
Date: 13th November 2018
Location: Governance and MDM Theatre
Time: 11:10 - 11:40
Organisation: TIBCO
About: The big data phenomenon continues to accelerate, resulting in multiple data lakes at most organisations. However, according to Gartner, “Through 2019, 90% of the information assets from big data analytic efforts will be siloed and unusable across multiple business processes.”
Are you ready to unleash this data from these silos and deliver the insights your organisation needs to drive compelling customer experiences, innovative new products and optimized operations? In this session you will learn how to apply data virtualisation to: - Access, transform and deliver data from across your lakes, clouds and other data sources - Empower a range of analytic users and tools with all the data they need - Move rapidly to a modern and flexible data architecture for the long run In addition, you will see a demonstration of data virtualisation in action.
Big Data LDN 2018: CONSISTENT SECURITY, GOVERNANCE AND FLEXIBILITY FOR ALL WO...Matt Stubbs
Date: 14th November 2018
Location: Data-Driven Ldn Theatre
Time: 12:30 - 13:00
Organisation: Cloudera
About: The growth of public cloud is reinforcing the need to think more carefully about taking a consistent approach to data governance as technology teams build out a flexible and agile infrastructure to meet the demands of the business.
Join this session to learn more about Cloudera's recommended approach for enterprise-grade security and governance and how to ensure a consistent framework across private, public and on-premises environments.
Big Data LDN 2018: MICROLISE: USING BIG DATA AND AI IN TRANSPORT AND LOGISTICSMatt Stubbs
Date: 14th November 2018
Location: Data-Driven Ldn Theatre
Time: 11:10 - 11:40
Organisation: Microlise
About: Microlise are a leading provider of technology solutions to the transport and logistics industry worldwide. Discover how, with over 400,000 connected assets generating billions of messages a day, Microlise is evolving its platform to bring real-time analytics to its customers to improve safety, security and efficiency outcomes.
Big Data LDN 2018: EXPERIAN: MAXIMISE EVERY OPPORTUNITY IN THE BIG DATA UNIVERSEMatt Stubbs
Date: 14th November 2018
Location: Data-Driven Ldn Theatre
Time: 10:30 - 11:00
Speaker: Anna Matty
Organisation: Experian
About: Today there is a widespread focus on the 'how' in relation to problem solving. How can we gain better knowledge of what consumers want, or need? How can we be more efficient, reduce the cost to serve, or grow the lifetime value of a customer? But, how do you move to a place where you are not only solving a problem, you are redesigning the entire strategic potential of that problem? You are being armed with insight on what the problem is.
Data and innovation offer huge potential to revolutionise all markets. There is an opportunity to be one step ahead of the need, to redesign journeys and enhance enterprise strategies. To do this you need access to the most advanced analytics but also the best quality, including variations and types of data, and then the technology that can act on this insight. Data science can present a unique opportunity for uncovered growth and accelerate your business through strategic innovation – fast. In this session you will hear more about how today's analytics can move from a single task, to an ongoing strategic opportunity. An opportunity that helps you move at the speed of the market and helps you maximise every opportunity.
Big Data LDN 2018: A LOOK INSIDE APPLIED MACHINE LEARNINGMatt Stubbs
Date: 13th November 2018
Location: Data-Driven Ldn Theatre
Time: 13:10 - 13:40
Speaker: Brian Goral
Organisation: Cloudera
About: The field of machine learning (ML) ranges from the very practical and pragmatic to the highly theoretical and abstract. This talk describes several of the challenges facing organisations that want to leverage more of their data through ML, including some examples of the applied algorithms that are already delivering value in business contexts.
Big Data LDN 2018: DEUTSCHE BANK: THE PATH TO AUTOMATION IN A HIGHLY REGULATE...Matt Stubbs
Date: 13th November 2018
Location: Data-Driven Ldn Theatre
Time: 12:30 - 13:00
Speaker: Paul Wilkinson, Naveen Gupta
Organisation: Cloudera
About: Investment banks are faced with some of the toughest regulatory requirements in the world. In a market where data is increasing and changing at extraordinary rates the journey with data governance never ends.
In this session, Deutsche Bank will share their journey with big data and explain some of the processes and techniques they have employed to prepare the bank for today’s challenges and tomorrow’s opportunities.
Brought to you by Naveen Gupta, VP Software Engineering, Deutsche Bank and Paul Wilkinson, Principal Solutions Architect, Cloudera.
Big Data LDN 2018: FROM PROLIFERATION TO PRODUCTIVITY: MACHINE LEARNING DATA ...Matt Stubbs
Date: 14th November 2018
Location: Self-Service Analytics Theatre
Time: 13:50 - 14:20
Speaker: Stephanie McReynolds
Organisation: Alation
About: Raw data is proliferating at an enormous rate. But so are our derived data assets - hundreds of dashboards, thousands of reports, millions of transformed data sets. Self-service analytics have ensured that this noise is making it increasingly hard to understand and trust data for decision-making. This trust gap is holding your organisation back from business outcomes.
European analytics leaders have found a way to close the gap between data and decision-making. From MunichRe to Pfizer and Daimler, analytics teams are adopting data catalogues for thousands of self-service analytics users.
Join us in this session to hear how data catalogues that activate data by incorporating machine learning can:
• Increase analyst productivity 20-40%
• Boost the understanding of the nuances of data and
• Establish trust in data-driven decisions with agile stewardship
Big Data LDN 2018: DATA APIS DON’T DISCRIMINATEMatt Stubbs
Date: 13th November 2018
Location: Self-Service Analytics Theatre
Time: 15:50 - 16:20
Speaker: Nishanth Kadiyala
Organisation: Progress
About: The exploding API economy, combined with an advanced analytics market projected to reach $30 billion by 2019, is forcing IT to expose more and more data through APIs. Business analysts, data engineers, and data scientists are still not happy because their needs never really made it into the existing API strategies. This is because most APIs are designed for application integration, but not for the data workers who are looking for APIs that facilitate direct data access to run complex analytics. Data APIs are specifically designed to provide that frictionless data access experience to support analytics across standard interoperable interfaces such as OData (REST) or ODBC/JDBC (SQL). Consider expanding your API strategy to service the developers with open analytics in this $30 billion market.
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).
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.
Techniques to optimize the pagerank algorithm usually fall in two categories. One is to try reducing the work per iteration, and the other is to try reducing the number of iterations. These goals are often at odds with one another. Skipping computation on vertices which have already converged has the potential to save iteration time. Skipping in-identical vertices, with the same in-links, helps reduce duplicate computations and thus could help reduce iteration time. Road networks often have chains which can be short-circuited before pagerank computation to improve performance. Final ranks of chain nodes can be easily calculated. This could reduce both the iteration time, and the number of iterations. If a graph has no dangling nodes, pagerank of each strongly connected component can be computed in topological order. This could help reduce the iteration time, no. of iterations, and also enable multi-iteration concurrency in pagerank computation. The combination of all of the above methods is the STICD algorithm. [sticd] For dynamic graphs, unchanged components whose ranks are unaffected can be skipped altogether.
Chatty Kathy - UNC Bootcamp Final Project Presentation - Final Version - 5.23...John Andrews
SlideShare Description for "Chatty Kathy - UNC Bootcamp Final Project Presentation"
Title: Chatty Kathy: Enhancing Physical Activity Among Older Adults
Description:
Discover how Chatty Kathy, an innovative project developed at the UNC Bootcamp, aims to tackle the challenge of low physical activity among older adults. Our AI-driven solution uses peer interaction to boost and sustain exercise levels, significantly improving health outcomes. This presentation covers our problem statement, the rationale behind Chatty Kathy, synthetic data and persona creation, model performance metrics, a visual demonstration of the project, and potential future developments. Join us for an insightful Q&A session to explore the potential of this groundbreaking project.
Project Team: Jay Requarth, Jana Avery, John Andrews, Dr. Dick Davis II, Nee Buntoum, Nam Yeongjin & Mat Nicholas