Building general purpose RL algorithms that can efficiently solve a wide variety of problems will require encoding the right structure and representations into our models. A key component of our generalization capability is our ability to develop an internal model of the world that can be used for robust prediction and efficient planning. In this talk I will discuss work that leverages...

Robots that can operate in an open, unstructured environment and perform a wide range of tasks have been a long-standing goal of artificial intelligence. In this talk, I will go over how we can build such robots using tools from deep learning and reinforcement learning. In particular, I will address some of the key bottlenecks to applying deep reinforcement learning to real world domains like...

As technology nodes scale, it becomes increasingly difficult to bring innovation to Application Specific Integrated Circuits (ASICs), as the complexity of design rules and the substantial impact of layout parasitics not only delays time-to-market, but is also more expensive. Emphasis is therefore placed on improving the ASIC design flow, with the primary focus of addressing inefficiencies in this...

With the end of Moore's Law and Dennard scaling, the continuous demand for higher computing performance and efficiency is increasingly met through specialization of digital processor implementations. In particular, numerical data processing and machine learning applications incur high computational costs but often have common computational structures, acting as prime targets for hardware...

Nanoscience Research Conference: Molecular Foundry 2021 User Meeting - Free & Online The Molecular Foundry is a nanoscience research facility at Berkeley Lab that provides visiting researchers (“users”) with access to cutting-edge expertise and instrumentation in a collaborative, multidisciplinary environment. The Molecular Foundry’s 2021 Annual User Meeting will be hosted virtually with...

Beginning with Robert Brown’s original observation in 1828, various techniques have been developed to study the hydrodynamics and interactions of particles in solution. These techniques have inspired or been followed by development of theories that are capable of describing these fundamental aspects of micron-scale particles. Yet, many of the underlying assumptions break down at the nanoscale...

The last decade has seen a dramatic capture of digital material for machine learning production. This data is the basis for sense-making in AI, not as classical representations of the world with individual meaning, but as mass collections: ground truth for machine abstractions and operations. OpenAI’s GPT-3 language model is trained on a corpus of 1 billion words, ImageNet contains over 14...

The last decade has seen a dramatic capture of digital material for machine learning production. This data is the basis for sense-making in AI, not as classical representations of the world with individual meaning, but as mass collections: ground truth for machine abstractions and operations. OpenAI’s GPT-3 language model is trained on a corpus of 1 billion words, ImageNet contains over 14...

The inaugural ACM conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO ’21) aims to highlight work where techniques from algorithms, optimization, and mechanism design, along with insights from other disciplines, can help improve equity and access to opportunity for historically disadvantaged and underserved communities. The conference is organized by the...

This seminar meeting will be done over zoom at this link: https://berkeley.zoom.us/j/91273049368?pwd=RFBMVzNtR3RDSjJRQ1F2TTdPanc5dz09 The atomic and molecular assembly processes on the surface of two-dimensional (2D) layered crystals have gained significant research interest for fabrication of well-ordered nanostructures and vertical heterostructures based on 2D crystals. The structure and...