A Berkeley View of Systems Challenges for AI

With the increasing commoditization of computer vision, speech recognition and machine translation systems and the widespread deployment of learning-based back-end technologies such as digital advertising and intelligent infrastructures, AI (Artificial Intelligence) has moved from research labs to production. These changes have been made possible by unprecedented levels of data and computation, by methodological advances in machine learning, by innovations in systems software and architectures, and by the broad accessibility of these technologies. The next generation of AI systems promises to accelerate these developments and increasingly impact our lives via frequent interactions and making (often mission-critical) decisions on our behalf, often in highly personalized contexts. Realizing this promise, however, raises daunting challenges. In particular, we need AI systems that make timely and safe decisions in unpredictable environments, that are robust against sophisticated adversaries, and that can process ever increasing amounts of data across organizations and individuals without compromising confidentiality. These challenges will be exacerbated by the end of the Moore's Law, which will constrain the amount of data these technologies can store and process. In this paper, we propose several open research directions in systems, architectures, and security that can address these challenges and help unlock AI's potential to improve lives and society.

[1]  William J. Bolosky,et al.  Mach: A New Kernel Foundation for UNIX Development , 1986, USENIX Summer.

[2]  Silvio Micali,et al.  How to play ANY mental game , 1987, STOC.

[3]  Avi Wigderson,et al.  Completeness theorems for non-cryptographic fault-tolerant distributed computation , 1988, STOC '88.

[4]  Michael McCloskey,et al.  Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem , 1989 .

[5]  Doron Rotem,et al.  Random Sampling from Database Files: A Survey , 1990, SSDBM.

[6]  Richard S. Sutton,et al.  Integrated Architectures for Learning, Planning, and Reacting Based on Approximating Dynamic Programming , 1990, ML.

[7]  Geoffrey E. Hinton,et al.  Feudal Reinforcement Learning , 1992, NIPS.

[8]  Satinder P. Singh,et al.  Reinforcement Learning with a Hierarchy of Abstract Models , 1992, AAAI.

[9]  Sebastian Thrun,et al.  Finding Structure in Reinforcement Learning , 1994, NIPS.

[10]  Gerald Tesauro,et al.  Temporal Difference Learning and TD-Gammon , 1995, J. Int. Comput. Games Assoc..

[11]  Jochen Liedtke,et al.  On micro-kernel construction , 1995, SOSP.

[12]  David D. Clark,et al.  The design philosophy of the DARPA internet protocols , 1988, SIGCOMM '88.

[13]  Stuart J. Russell,et al.  Reinforcement Learning with Hierarchies of Machines , 1997, NIPS.

[14]  Helen J. Wang,et al.  Online aggregation , 1997, SIGMOD '97.

[15]  Thomas G. Dietterich.  The MAXQ Method for Hierarchical Reinforcement Learning , 1998, ICML.

[16]  Sebastian Thrun,et al.  Lifelong Learning Algorithms , 1998, Learning to Learn.

[17]  David Saad,et al.  On-Line Learning in Neural Networks , 1999 .

[18]  Doina Precup,et al.  Between MDPs and Semi-MDPs: A Framework for Temporal Abstraction in Reinforcement Learning , 1999, Artif. Intell..

[19]  Howard Gobioff,et al.  The Google file system , 2003, SOSP '03.

[20]  Benny Pinkas,et al.  Fairplay - Secure Two-Party Computation System , 2004, USENIX Security Symposium.

[21]  H. Sebastian Seung,et al.  Learning to Walk in 20 Minutes , 2005 .

[22]  Yuan Yu,et al.  Dryad: distributed data-parallel programs from sequential building blocks , 2007, EuroSys '07.

[23]  Cynthia Dwork,et al.  Differential Privacy: A Survey of Results , 2008, TAMC.

[24]  Stefan Schaal,et al.  2008 Special Issue: Reinforcement learning of motor skills with policy gradients , 2008 .

[25]  Petros Drineas,et al.  CUR matrix decompositions for improved data analysis , 2009, Proceedings of the National Academy of Sciences.

[26]  Luiz André Barroso,et al.  The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines , 2009, The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines.

[27]  Rajat Raina,et al.  Large-scale deep unsupervised learning using graphics processors , 2009, ICML '09.

[28]  Moni Naor,et al.  Differential privacy under continual observation , 2010, STOC '10.

[29]  Léon Bottou,et al.  Large-Scale Machine Learning with Stochastic Gradient Descent , 2010, COMPSTAT.

[30]  Stephen J. Wright,et al.  Hogwild: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent , 2011, NIPS.

[31]  Steven Hand,et al.  CIEL: A Universal Execution Engine for Distributed Data-Flow Computing , 2011, NSDI.

[32]  Elaine Shi,et al.  Private and Continual Release of Statistics , 2010, TSEC.

[33]  Marc'Aurelio Ranzato,et al.  Large Scale Distributed Deep Networks , 2012, NIPS.

[34]  Kun Li,et al.  The MADlib Analytics Library or MAD Skills, the SQL , 2012, Proc. VLDB Endow..

[35]  Michael J. Franklin,et al.  Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster Computing , 2012, NSDI.

[36]  Joseph Gonzalez,et al.  PowerGraph: Distributed Graph-Parallel Computation on Natural Graphs , 2012, OSDI.

[37]  Ion Stoica,et al.  BlinkDB: queries with bounded errors and bounded response times on very large data , 2012, EuroSys '13.

[38]  Qiang Yang,et al.  Lifelong Machine Learning Systems: Beyond Learning Algorithms , 2013, AAAI Spring Symposium: Lifelong Machine Learning.

[39]  Stephane Ross,et al.  Interactive Learning for Sequential Decisions and Predictions , 2013 .

[40]  Ittai Anati,et al.  Innovative Technology for CPU Based Attestation and Sealing , 2013 .

[41]  Tim Kraska,et al.  MLI: An API for Distributed Machine Learning , 2013, 2013 IEEE 13th International Conference on Data Mining.

[42]  Alexander J. Smola,et al.  Scaling Distributed Machine Learning with the Parameter Server , 2014, OSDI.

[43]  Galen C. Hunt,et al.  Shielding Applications from an Untrusted Cloud with Haven , 2014, OSDI.

[44]  Trevor Darrell,et al.  Deep Domain Confusion: Maximizing for Domain Invariance , 2014, CVPR 2014.

[45]  Thomas Hofmann,et al.  Communication-Efficient Distributed Dual Coordinate Ascent , 2014, NIPS.

[46]  Trevor Darrell,et al.  Caffe: Convolutional Architecture for Fast Feature Embedding , 2014, ACM Multimedia.

[47]  Joan Bruna,et al.  Intriguing properties of neural networks , 2013, ICLR.

[48]  Claudia Eckert,et al.  Is Feature Selection Secure against Training Data Poisoning? , 2015, ICML.

[49]  Xiaojin Zhu,et al.  The Security of Latent Dirichlet Allocation , 2015, AISTATS.

[50]  Christos Gkantsidis,et al.  VC3: Trustworthy Data Analytics in the Cloud Using SGX , 2015, 2015 IEEE Symposium on Security and Privacy.

[51]  Zheng Zhang,et al.  MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems , 2015, ArXiv.

[52]  Sergey Levine,et al.  Trust Region Policy Optimization , 2015, ICML.

[53]  Pieter Abbeel,et al.  Image Object Label 3 D CAD Model Candidate Grasps Google Object Recognition Engine Google Cloud Storage Select Feasible Grasp with Highest Success Probability Pose EstimationCamera Robots Cloud 3 D Sensor , 2014 .

[54]  Xiaojin Zhu,et al.  Using Machine Teaching to Identify Optimal Training-Set Attacks on Machine Learners , 2015, AAAI.

[55]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[56]  Jonathon Shlens,et al.  Explaining and Harnessing Adversarial Examples , 2014, ICLR.

[57]  Xinlei Chen,et al.  Never-Ending Learning , 2012, ECAI.

[58]  Somesh Jha,et al.  Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures , 2015, CCS.

[59]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[60]  M. Schatz,et al.  Big Data: Astronomical or Genomical? , 2015, PLoS biology.

[61]  Asterios Katsifodimos,et al.  Apache Flink: Stream Analytics at Scale , 2016, 2016 IEEE International Conference on Cloud Engineering Workshop (IC2EW).

[62]  Ameet Talwalkar,et al.  MLlib: Machine Learning in Apache Spark , 2015, J. Mach. Learn. Res..

[63]  Saman P. Amarasinghe,et al.  Weld : A Common Runtime for High Performance Data Analytics , 2016 .

[64]  Heike Freud,et al.  On Line Learning In Neural Networks , 2016 .

[65]  David M. Eyers,et al.  SCONE: Secure Linux Containers with Intel SGX , 2016, OSDI.

[66]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[67]  Kai-Fu Tang,et al.  Inquire and Diagnose : Neural Symptom Checking Ensemble using Deep Reinforcement Learning , 2016 .

[68]  Sergey Levine,et al.  End-to-End Training of Deep Visuomotor Policies , 2015, J. Mach. Learn. Res..

[69]  Xin Wang,et al.  Clipper: A Low-Latency Online Prediction Serving System , 2016, NSDI.

[70]  David A. Patterson,et al.  In-datacenter performance analysis of a tensor processing unit , 2017, 2017 ACM/IEEE 44th Annual International Symposium on Computer Architecture (ISCA).

[71]  Paramvir Bahl,et al.  Live Video Analytics at Scale with Approximation and Delay-Tolerance , 2017, NSDI.

[72]  Wojciech Zaremba,et al.  Domain randomization for transferring deep neural networks from simulation to the real world , 2017, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[73]  Blaise Agüera y Arcas,et al.  Communication-Efficient Learning of Deep Networks from Decentralized Data , 2016, AISTATS.

[74]  Ion Stoica,et al.  DDCO: Discovery of Deep Continuous Options for Robot Learning from Demonstrations , 2017, CoRL.

[75]  Vitaly Shmatikov,et al.  Membership Inference Attacks Against Machine Learning Models , 2016, 2017 IEEE Symposium on Security and Privacy (SP).

[76]  Anne E Carpenter,et al.  Opportunities and obstacles for deep learning in biology and medicine , 2017, bioRxiv.