Intel 20180608 v2
Download as PPTX, PDF1 like298 views
This document summarizes a presentation about the future of AI and Fabric for Deep Learning (FfDL). It discusses how deep learning has advanced due to increased data and computing power, but that commonsense reasoning will require more research. FfDL is introduced as an open source project that aims to make deep learning accessible and scalable across frameworks. It uses a microservices architecture on Kubernetes to manage training jobs efficiently. Research is ongoing to further develop explainable and robust AI capabilities.
Intel 20180608 v2
- 1. Future of AI & FfDL Jim Spohrer (IBM) and Animesh Singh (IBM) http://slideshare.net/spohrer/intel_20180608_v2 June 8, 2018 - Skype Intel Skype PresentationIntel Hosts: John Miranda and Michael Jacobson 6/8/2018 IBM #OpenTechAI 1
- 2. IBM Contacts 6/8/2018 IBM #OpenTechAI 2 Jim Spohrer <spohrer@us.ibm.com> IBM Research – Almaden San Jose, CA Animesh Singh <singhan@us.ibm.com> IBM Silicon Valley Lab San Jose, DC Vijay Bommireddipalli <vijayrb@us.ibm.com> CODAIT, San Francisco, CACenter
- 3. 6/8/2018 © IBM UPWard 2016 3 AI (Artificial Intelligence) is popular again… you see it mentioned on billboards in SF However, pattern recognition does not equal AI Deep learning works if you have lots of data and compute power We finally have lots of data and compute power – hurray!!! So finally, deep learning for pattern recognition is working pretty well However, AI is more than deep learning for pattern recognition… AI requires commonsense reasoning – that will take another 5-10 years of research How do we know this? Look at the AI leaderboards – we will get to that…
- 4. Future of AI 6/8/2018 (c) IBM 2017, Cognitive Opentech Group 4 … when will your smartphone be able to take and pass any online course? And then be your coach, so you can pass too?
- 5. 6/8/2018 Understanding Cognitive Systems 5
- 6. Future of AI 6/8/2018 © IBM Cognitive Opentech Group (COG) 6 Dota 2 “Deep Learning” for “AI Pattern Recognition” depends on massive amounts of “labeled data” and computing power available since ~2012; Labeled data is simply input and output pairs, such as a sound and word, or image and word, or English sentence and French sentence, or road scene and car control settings – labeled data means having both input and output data in massive quantities. For example, 100K images of skin, half with skin cancer and half without to learn to recognize presence of skin cancer.
- 7. Every 20 years, compute costs are down by 1000x • Cost of Digital Workers • Moore’s Law can be thought of as lowering costs by a factor of a… • Thousand times lower in 20 years • Million times lower in 40 years • Billion times lower in 60 years • Smarter Tools (Terascale) • Terascale (2017) = $3K • Terascale (2020) = ~$1K • Narrow Worker (Petascale) • Recognition (Fast) • Petascale (2040) = ~$1K • Broad Worker (Exascale) • Reasoning (Slow) • Exascale (2060) = ~$1K 76/8/2018 (c) IBM 2017, Cognitive Opentech Group 2080204020001960 $1K $1M $1B $1T 206020201980 +/- 10 years $1 Person Average Annual Salary (Living Income) Super Computer Cost Mainframe Cost Smartphone Cost T P E T P E AI Progress on Open Leaderboards Benchmark Roadmap to solve AI/IA
- 8. GDP/Employee 6/8/2018 (c) IBM 2017, Cognitive Opentech Group 8 (Source) Lower compute costs translate into increasing productivity and GDP/employees for nations Increasing productivity and GDP/employees should translate into wealthier citizens AI Progress on Open Leaderboards Benchmark Roadmap to solve AI/IA
- 9. Leaderboards Framework AI Progress on Open Leaderboards - Benchmark Roadmap Perceive World Develop Cognition Build Relationships Fill Roles Pattern recognition Video understanding Memory Reasoning Social interactions Fluent conversation Assistant & Collaborator Coach & Mediator Speech Actions Declarative Deduction Scripts Speech Acts Tasks Institutions Chime Thumos SQuAD SAT ROC Story ConvAI Images Context Episodic Induction Plans Intentions Summarizatio n Values ImageNet VQA DSTC RALI General-AI Translation Narration Dynamic Abductive Goals Cultures Debate Negotiation WMT DeepVideo Alexa Prize ICCMA AT Learning from Labeled Training Data and Searching (Optimization) Learning by Watching and Reading (Education) Learning by Doing and being Responsible (Exploration) 2015 2018 2021 2024 2027 2030 2033 2036 6/8/2018 (c) IBM 2017, Cognitive Opentech Group 9 Which experts would be really surprised if it takes less time… and which experts really surprised if it takes longer? Approx. Year Human Level ->
- 10. 6/8/2018 10 1955 1975 1995 2015 2035 2055 Better Building Blocks
- 11. Build: 10 million minutes of experience 6/8/2018 Understanding Cognitive Systems 11
- 12. Build: 2 million minutes of experience 6/8/2018 Understanding Cognitive Systems 12
- 13. Build: Hardware < Software < Data < Experience 6/8/2018 Understanding Cognitive Systems 13
- 14. Types: Progression of models and capabilities 6/8/2018 Understanding Cognitive Systems 14 Task & World Model/ Planning & Decisions Self Model/ Capacity & Limits User Model/ Episodic Memory Institutions Model/ Trust & Social Acts Tool + - - - Assistant ++ + - - Collaborator +++ ++ + - Coach ++++ +++ ++ + Mediator +++++ ++++ +++ ++ Cognitive Tool Cognitive Assistant Cognitive Collaborator Cognitive Coach Cognitive Mediator
- 15. “The best way to predict the future is to inspire the next generation of students to build it better” Digital Natives Transportation Water Manufacturing Energy Construction ICT Retail Finance Healthcare Education Government
- 16. 6/8/2018 IBM #OpenTechAI 16
- 17. Step Comment GitHub Get an account and read the guide Learn 3 R's - Read, Redo, Report Read (Medium/arXiv), Redo (GitHub), Report (Jupyter Notebook) Kaggle Compete in a Kaggle competition Leaderboards Compete to advance AI progress Design New Challenges build an AI system that can take and pass any online course, then switch to tutor-mode and help you pass Open Source Guide Establish open source culture in your organization 6/8/2018 IBM #OpenTechAI 17
- 18. Fabric for Deep Learning FfDL FfDL Github Page https://github.com/IBM/FfDL FfDL dwOpen Page https://developer.ibm.com/code/open/projects/fabri c-for-deep-learning-ffdl/ FfDL Announcement Blog http://developer.ibm.com/code/2018/03/20/fabric- for-deep-learning FfDL Technical Architecture Blog http://developer.ibm.com/code/2018/03/20/democr atize-ai-with-fabric-for-deep-learning Deep Learning as a Service within Watson Studio https://www.ibm.com/cloud/deep-learning Research paper: “Scalable Multi-Framework Management of Deep Learning Training Jobs” http://learningsys.org/nips17/assets/papers/paper_ 29.pdf FfDL 18 https://github.com/IBM/FfDL
- 19. …that automate decisions. …to build models…Use data… The Enterprise AI Process 19 Gather Data Analyze Data Machine Learning Deep Learning Deploy Model Maintain Model
- 20. Center for Open Source Data and AI Technologies March 30 2018 / © 2018 IBM Corporation codait (French) = coder/coded https://m.interglot.com/fr/en/codaitCode - Build and improve practical frameworks to enable more developers to realize immediate value (e.g. FfDL, Tensorflow Jupyter, Spark) Content – Showcase solutions to complex and real world AI problems Community – Bring developers and data scientists together to engage (e.g. MAX) Improving Enterprise AI lifecycle in Open Source Gather Data Analyze Data Machine Learning Deep Learning Deploy Model Maintain Model Python Data Science Stack Fabric for Deep Learning (FfDL) Mleap + PFA Scikit-LearnPandas Apache Spark Apache Spark Jupyter Model Asset eXchange Keras + Tensorflow CODAIT codait.org 20
- 21. Fabric for Deep Learning https://github.com/IBM/FfDL FfDL provides a scalable, resilient, and fault tolerant deep-learning framework FfDL Github Page https://github.com/IBM/FfDL FfDL dwOpen Page https://developer.ibm.com/code/open/projects/fabri c-for-deep-learning-ffdl/ FfDL Announcement Blog http://developer.ibm.com/code/2018/03/20/fabric- for-deep-learning FfDL Technical Architecture Blog http://developer.ibm.com/code/2018/03/20/democr atize-ai-with-fabric-for-deep-learning Deep Learning as a Service within Watson Studio https://www.ibm.com/cloud/deep-learning Research paper: “Scalable Multi-Framework Management of Deep Learning Training Jobs” http://learningsys.org/nips17/assets/papers/paper_ 29.pdf • Fabric for Deep Learning or FfDL (pronounced as ‘fiddle’) is an open source project which aims at making Deep Learning easily accessible to the people it matters the most i.e. Data Scientists, and AI developers. • FfDL Provides a consistent way to deploy, train and visualize Deep Learning jobs across multiple frameworks like TensorFlow, Caffe, PyTorch, Keras etc. • FfDL is being developed in close collaboration with IBM Research and IBM Watson. It forms the core of Watson`s Deep Learning service in open source. FfDL 21
- 22. Fabric for Deep Learning https://github.com/IBM/FfDL FfDL is built using Microservices architecture on Kubernetes • FfDL platform uses a microservices architecture to offer resilience, scalability, multi-tenancy, and security without modifying the deep learning frameworks, and with no or minimal changes to model code. • FfDL control plane microservices are deployed as pods on Kubernetes to manage this cluster of GPU- and CPU-enabled machines effectively • Tested Platforms: Minikube, IBM Cloud Public, IBM Cloud Private, GPUs using both Kubernetes feature gate Accelerators and NVidia device plugins 22
- 23. source code training definition Access to elastic compute leveraging Kubernetes Auto-allocation means infrastructure is used only when needed Kubernetes container training artifacts compute cluster NVIDIA Tesla K80, P100, V100 Cloud Object Storage Training assets are managed and tracked. IBM Cloud / Watson and Cloud Platform / © 2018 IBM Corporation 23
- 24. NVIDIA GPUs Kubernetes container orchestration training runs containers Model training distributed across containers server cluster dataset Cloud Object Storage 24
- 27. 27 FfDL: Research Papers http://learningsys.org/nips17/assets/papers/paper_29.pdf
- 28. And we offer more Model Asset Exchange MAX and Adversarial Robustness Toolbox ART 28
- 29. IBM Model Asset eXchange MAX MAX is a one stop exchange to find ML/DL models created using popular Machine Learning engines and provides a standardized approach to consume these models for training and inferencing. 29 developer.ibm.com/code/exchanges/models/
- 30. IBM Adversarial Robustness Toolbox ART ART is a library dedicated to adversarial machine learning. Its purpose is to allow rapid crafting and analysis of attacks and defense methods for machine learning models. The Adversarial Robustness Toolbox provides an implementation for many state-of-the-art methods for attacking and defending classifiers. 30 https://developer.ibm.com/code/open/projects/adver sarial-robustness-toolbox/ The Adversarial Robustness Toolbox contains implementations of the following attacks: Deep Fool (Moosavi-Dezfooli et al., 2015) Fast Gradient Method (Goodfellow et al., 2014) Jacobian Saliency Map (Papernot et al., 2016) Universal Perturbation (Moosavi-Dezfooli et al., 2016) Virtual Adversarial Method (Moosavi-Dezfooli et al., 2015) C&W Attack (Carlini and Wagner, 2016) NewtonFool (Jang et al., 2017) The following defense methods are also supported: Feature squeezing (Xu et al., 2017) Spatial smoothing (Xu et al., 2017) Label smoothing (Warde-Farley and Goodfellow, 2016) Adversarial training (Szegedy et al., 2013) Virtual adversarial training (Miyato et al., 2017)
- 31. FfDL Core of Deep Learning as a Service in Watson Studio 31
- 32. Model Lifecycle Management Machine Learning Runtimes Deep Learning Runtimes Authoring Tools Cloud Infrastructure as a Service • Most popular open source frameworks • IBM best-in-class frameworks • Create, collaborate, deploy, and monitor • Best of breed open source & IBM tools • Code (R, Python or Scala) and no-code/visual modeling tools • Fully managed service • Container-based resource management • Elastic pay as you go cpu/gpu power Watson Studio Tools for supporting the end-to-end AI workflow
- 33. 3 Train neural networks in parallel across NVIDIA GPUs. Pay only for what you use. Auto- deallocation means no more remembering to shutdown your cloud training instances. Monitor batch training experiments then compare cross-model performance without worrying about log transfers and scripts to visualize results. You focus on designing your neural networks. We’ll manage and track your assets. Python client, command line interface (CLI) or UI? You choose the tooling that best fits your existing workflows. Training history and assets are tracked then automatically transferred to the customer’s Object Storage for quick access. Deploy models into production then monitor them to evaluate performance. Capture new data for continuous learning and retrain models so they continually adapt to changing conditions. Deep Learning as a Service within Watson Studio Using FfDL as core
- 34. Neural Network Modeller within Watson Studio An intuitive drag-and-drop, no-code interface for designing neural network structure
- 35. DLaaS Training Dashboard in Watson Studio
- 36. 6/8/2018 © IBM 2015, IBM Upward University Programs Worldwide accelerating regional development 36
- 37. Trust: Two Communities6/8/2018 © IBM MAP COG2018 37 Service Science OpenTech AI Trust: Value Co-Creation, Transdisciplinary Trust: Ethical, Safe, Explainable, Open Communities Special Issue AI Magazine? Handbook of OpenTech AI?
- 38. Resilience: Rapidly Rebuilding From Scratch Dartnell L (2012) The Knowledge: How to Rebuild Civilization in the Aftermath of a Cataclysm. Westminster London: Penguin Books. 6/8/2018 © IBM MAP COG2018 38
Editor's Notes
- #2: Please reuse – contact spohrer@us.ibm.com Reference: Spohrer, J (2018) Future of AI & CODAIT/FfDL. Intel Skype Talk. June 8, 2018 URL Slides: http://slideshare.net/spohrer/intel_20180608_v2
- #3: Jim Spohrer Linkedin: https://www.linkedin.com/in/spohrer/ IBM: http://service-science.info/archives/4679 Vijay Bommireddipalli LinkedIn: https://www.linkedin.com/in/vijay-bommireddipalli-6b6b92/ IBM: http://codiat.org Frank Stein Linkedin: https://www.linkedin.com/in/frank-stein-8b67781b/ IBM: https://www.ibm.com/industries/federal/analytics
- #5: URL: https://www.ted.com/talks/noriko_arai_can_a_robot_pass_a_university_entrance_exam
- #7: 1950 Nathaniel Rochester (IBM) 701 first commercial computer that did super-human levels of numeric calculations routinely. He worked at MIT on arithmetic unit of WhirlWind I programmable computer. Dota 2 is most recent August 11, 2017 as a super-human game player in Valve Dota 2 competition – Elon Musk’s OpenAI result. Miles Bundage tracks gaming progress: http://www.milesbrundage.com/blog-posts/my-ai-forecasts-past-present-and-future-main-post DOTA2: https://blog.openai.com/more-on-dota-2/
- #8: What is beyond Exascale? Zetta (21), Yotta (24) Time dimension (x-axis) is plus or minus 10 years…. Daniel Pakkala (VTT) URL: https://aiimpacts.org/preliminary-prices-for-human-level-hardware/ Dan Gruhl: https://www.washingtonpost.com/archive/business/1983/11/06/in-pursuit-of-the-10-gigaflop-machine/012c995a-2b16-470b-96df-d823c245306e/?utm_term=.d4bde5652826 In 1983 10 GF was ~10 million. That's 24.55 million in today's dollars. or 2.4 billion for 1 TF in 1983 Today 1 TF is about $3k http://www.popsci.com/intel-teraflop-chip
- #9: Source: http://service-science.info/archives/4741
- #10: Expert predictions on HMLI: URL https://arxiv.org/pdf/1705.08807.pdf 2015 Pattern Recognition Speech: URL: http://spandh.dcs.shef.ac.uk/chime_challenge/chime2016/results.html 2015 Pattern Recognition Images: URL: http://www.image-net.org/ 2015 Patten Recognition Translation: URL: http://www.statmt.org/wmt17/ 2018 Video Understanding Actions: URL: http://www.thumos.info/home.html > Also UCF101 http://crcv.ucf.edu/data/UCF101.php 2018 Video Understanding Context: URL: http://visualqa.org/challenge.html 2018 Video Understanding DeepVideo: URL: http://cs.stanford.edu/people/karpathy/deepvideo/ 2021 Memory Declarative: URL: https://rajpurkar.github.io/SQuAD-explorer/ Also Allen AI Kaggle Science Challenge https://www.kaggle.com/c/the-allen-ai-science-challenge 2024 Reasoning Deduction: URL: http://www.satcompetition.org/ 2027: Social Interaction Scripts: URL: https://competitions.codalab.org/competitions/15333 2030: Fluent Conversation Speech Acts: URL: http://convai.io/ 2030: Fluent Conversation Intentions: URL: http://workshop.colips.org/dstc6/ 2030: Fluent Conversation Alexa Prize: URL: https://developer.amazon.com/alexaprize 2033: Assistant & Collaborator Summarization: URL: http://rali.iro.umontreal.ca/rali/?q=en/Automatic%20summarization 2033: Assistant & Collaborator Debate: URL: http://argumentationcompetition.org/2015/ 2036: Coach & Mediator General AI: URL: https://www.general-ai-challenge.org/ 2036: Coach & Mediator Negotiation: URL: https://easychair.org/cfp/AT2017
- #11: The weakest link is what needs to be improved – according to system scientists. Accessing help, service, experts is the weakest link in most systems. By 2035 the phone may have the power of one human brain – by 2055 the phone may have the power of all human brains. Before trying to answer the question about which types of sciences are more important – the ones that try to explain the external world or the ones that try to explain the internal world – consider this, slide that shows the different telephones that I have used in my life. I grew up in rural Maine, where we had a party line telephone because we were somewhat remote on our farm in Newburgh, Maine. However, over the years phones got much better…. So in 2035 or 2055, who are you going to call when you need help?
- #14: Where is the variety? Hardware and even software standardizing into modules and algorithms…. Data will standardize next into categories and types…. Experience is where the uniqueness is, and variety and variability, and identity.
- #16: By 2036, there will be an accumulation of knowledge as well as a distribution of knowledge in service systems globally. We need to ensure as there is knowledge accumulation that service systems at all scale become more resilient. Leading to the capability of rapid rebuilding of service systems across scales, by T-shaped people who understand how to rapidly rebuild – knowledge has been chunked, modularized, and put into networks that support rapid rebuilding.
- #17: Source: Vijay Bommireddipally (CODAIT Director) and Fred Reiss (CODAIT Chief Architect)
- #39: URL Amazon: https://www.amazon.com/Knowledge-Rebuild-Civilization-Aftermath-Cataclysm-ebook/dp/B00DMCV5YS/ URL TED Talk: https://www.youtube.com/watch?v=CdTzsbqQyhY Citation: Dartnell L (2012) The Knowledge: How to Rebuild Civilization in the Aftermath of a Cataclysm. Westminster London: Penguin Books. Jim Spohrer Blogs: Grand Challenge: http://service-science.info/archives/2189 Re-readings: http://service-science.info/archives/4416