A Berkeley View of Systems Challenges for AI

By Ion Stoica, Dawn Song, Raluca Ada Popa, David A. Patterson, Michael W. Mahoney, Randy Howard Katz, Anthony D. Joseph, Michael I. Jordan, Joseph M Hellerstein, Joseph E. Gonzalez and 4 more. 2017

213 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Information Systems and Management

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems and Management

Shane Legg

Citations in the corpus

STRIP: a defence against trojan attacks on deep neural networks (2019, background): “However, the safety of ML system deployments has now become a realistic security concern [4], [5].”
Deep Learning: A Critical Appraisal (2018, background): “A partial list includes Brenden Lake and Marco Baroni (2017), François Chollet (2017), Robin Jia and Percy Liang (2017), Dileep George and others at Vicarious (Kansky et al., 2017) and Pieter Abbeel and colleagues at Berkeley (Stoica et al., 2017).”
LAG: Lazily Aggregated Gradient for Communication-Efficient Distributed Learning (2018, background): “We are particularly interested in the scenarios where communication between the central server and the local workers is costly, as is the case with the Federated Learning paradigm (McMahan et al., 2017; Smith et al., 2017), and the cloud-edge AI systems (Stoica et al., 2017).”
AI Safety Gridworlds (2017, background): “These long-term safety challenges might not be as relevant before we build powerful general AI systems, and are complementary to the short-term safety challenges of deploying today’s systems (Stoica et al., 2017).”
MLCapsule: Guarded Offline Deployment of Machine Learning as a Service (2018, background): “We believe that this is an important step towards the overall vision of data privacy in machine learning [26] as well as secure ML and AI [32].”
Towards Safe Artificial General Intelligence (2018, background): “More up-to-date references are provided by this chapter, and by a number of recent research agendas and problem collections (Amodei, Olah, et al., 2016; Leike, Martic, et al., 2017; Russell, Dewey, et al., 2016; Soares and Fallenstein, 2017; Stoica et al., 2017; Taylor et al., 2016).”
Federated Learning via Over-the-Air Computation (2018, background): “2961673 the recent breakthroughs in machine learning (ML) techniques particularly deep learning, as well as the unprecedented levels of computing power [2].”
AGI Safety Literature Review (2018, methodology): “GI problems down to concrete machine learning problems: Amodei, Olah, et al. (2016) at OpenAI et al., Leike, Martic, et al. (2017) at DeepMind, and Taylor et al. (2016) also at MIRI. In the agenda by Stoica et al. (2017) at UC Berkeley, the connection to AGI has all but vanished. For brevity, we wil”
Learning with Humans in the Loop (2020, background): “for an overview of challenges and approaches to dealing with the systems side of the above algorithms [61].”
Edge Intelligence: Paving the Last Mile of Artificial Intelligence With Edge Computing (2019, background): “In December 2017, in a white paper, “A Berkeley View of Systems Challenges for AI” [39] published by the University of California Berkeley, the cloud–edge AI system is envisioned as an important research direction to achieve the goal of mission-critical and personalized AI.”
FutureMapping: The Computational Structure of Spatial AI Systems (2018, background): “In the recent technical report from leading UC Berkeley researchers on systems challenges for AI [51], it is argued that devices which can act intelligently in their environments via continual learning must be capable of Simulated Reality (SR) which can “faithfully simulate the real-world…”