arXiv subjects

Jesse Zhang

Publications and source records attributed to Jesse Zhang.

28 records · Page 2Linked to original sources

arXiv2022

Learning to Synthesize Programs as Interpretable and Generalizable Policies

Recently, deep reinforcement learning (DRL) methods have achieved impressive performance on tasks in a variety of domains. However, neural network policies produced with DRL methods are not human-interpretable and often have difficulty generalizing to novel scenarios. To address these issues, prior works explore learning programmatic policies that are more interpretable and structured for generalization. Yet, these works either employ limited policy representations (e.g. decision trees, state machines, or predefined program templates) or require stronger supervision (e.g. input/output state pairs or expert demonstrations). We present a framework that instead learns to synthesize a program, which details the procedure to solve a task in a flexible and expressive manner, solely from reward signals. To alleviate the difficulty of learning to compose programs to induce the desired agent behavior from scratch, we propose to first learn a program embedding space that continuously parameterizes diverse behaviors in an unsupervised manner and then search over the learned program embedding space to yield a program that maximizes the return for a given task. Experimental results demonstrate that the proposed framework not only learns to reliably synthesize task-solving programs but also outperforms DRL and program synthesis baselines while producing interpretable and more generalizable policies. We also justify the necessity of the proposed two-stage learning scheme as well as analyze various methods for learning the program embedding.

cs.LG↗

arXiv2021

Stable High Power deep-UV Enhancement Cavity in Ultra High Vacuum with Fluoride Coatings

We demonstrate the superior performance of fluoride coated versus oxide coated mirrors in long term vacuum operation of a high power deep-ultraviolet enhancement cavity. In high vacuum ($10^{-8}$ mbar), the fluoride optics can maintain up to a record-high 10 w of stable intracavity power on one hour time scales, whereas for the oxide optics, we observe rapid degradation at lower intracavity powers with a rate that increases with power. After observing degradation in high vacuum, we can recover the fluoride and oxide optics with oxygen; however, this recovery process becomes ineffective after several applications. For fluoride coatings, we see that initial UV conditioning in an oxygen environment helps to improve the performances of the optics. In oxygen-rich environments from $\sim 10^{-4}$ mbar to 1 mbar, the fluoride optics can stably maintain up to 20 W of intracavity power on several-hour time scales whereas for the oxide optics there is immediate degradation with a rate that increases with decreasing oxygen pressure.

physics.optics↗

arXiv2021

Greenplum: A Hybrid Database for Transactional and Analytical Workloads

Demand for enterprise data warehouse solutions to support real-time Online Transaction Processing (OLTP) queries as well as long-running Online Analytical Processing (OLAP) workloads is growing. Greenplum database is traditionally known as an OLAP data warehouse system with limited ability to process OLTP workloads. In this paper, we augment Greenplum into a hybrid system to serve both OLTP and OLAP workloads. The challenge we address here is to achieve this goal while maintaining the ACID properties with minimal performance overhead. In this effort, we identify the engineering and performance bottlenecks such as the under-performing restrictive locking and the two-phase commit protocol. Next we solve the resource contention issues between transactional and analytical queries. We propose a global deadlock detector to increase the concurrency of query processing. When transactions that update data are guaranteed to reside on exactly one segment we introduce one-phase commit to speed up query processing. Our resource group model introduces the capability to separate OLAP and OLTP workloads into more suitable query processing mode. Our experimental evaluation on the TPC-B and CH-benCHmark benchmarks demonstrates the effectiveness of our approach in boosting the OLTP performance without sacrificing the OLAP performance.

cs.DB↗

arXiv2021

TRECVID 2020: A comprehensive campaign for evaluating video retrieval tasks across multiple application domains

The TREC Video Retrieval Evaluation (TRECVID) is a TREC-style video analysis and retrieval evaluation with the goal of promoting progress in research and development of content-based exploitation and retrieval of information from digital video via open, metrics-based evaluation. Over the last twenty years this effort has yielded a better understanding of how systems can effectively accomplish such processing and how one can reliably benchmark their performance. TRECVID has been funded by NIST (National Institute of Standards and Technology) and other US government agencies. In addition, many organizations and individuals worldwide contribute significant time and effort. TRECVID 2020 represented a continuation of four tasks and the addition of two new tasks. In total, 29 teams from various research organizations worldwide completed one or more of the following six tasks: 1. Ad-hoc Video Search (AVS), 2. Instance Search (INS), 3. Disaster Scene Description and Indexing (DSDI), 4. Video to Text Description (VTT), 5. Activities in Extended Video (ActEV), 6. Video Summarization (VSUM). This paper is an introduction to the evaluation framework, tasks, data, and measures used in the evaluation campaign.

cs.CV↗

arXiv2020

COG: Connecting New Skills to Past Experience with Offline Reinforcement Learning

Reinforcement learning has been applied to a wide variety of robotics problems, but most of such applications involve collecting data from scratch for each new task. Since the amount of robot data we can collect for any single task is limited by time and cost considerations, the learned behavior is typically narrow: the policy can only execute the task in a handful of scenarios that it was trained on. What if there was a way to incorporate a large amount of prior data, either from previously solved tasks or from unsupervised or undirected environment interaction, to extend and generalize learned behaviors? While most prior work on extending robotic skills using pre-collected data focuses on building explicit hierarchies or skill decompositions, we show in this paper that we can reuse prior data to extend new skills simply through dynamic programming. We show that even when the prior data does not actually succeed at solving the new task, it can still be utilized for learning a better policy, by providing the agent with a broader understanding of the mechanics of its environment. We demonstrate the effectiveness of our approach by chaining together several behaviors seen in prior datasets for solving a new task, with our hardest experimental setting involving composing four robotic skills in a row: picking, placing, drawer opening, and grasping, where a +1/0 sparse reward is provided only on task completion. We train our policies in an end-to-end fashion, mapping high-dimensional image observations to low-level robot control commands, and present results in both simulated and real world domains. Additional materials and source code can be found on our project website: https://sites.google.com/view/cog-rl

cs.LG↗

arXiv2020

TRECVID 2019: An Evaluation Campaign to Benchmark Video Activity Detection, Video Captioning and Matching, and Video Search & Retrieval

The TREC Video Retrieval Evaluation (TRECVID) 2019 was a TREC-style video analysis and retrieval evaluation, the goal of which remains to promote progress in research and development of content-based exploitation and retrieval of information from digital video via open, metrics-based evaluation. Over the last nineteen years this effort has yielded a better understanding of how systems can effectively accomplish such processing and how one can reliably benchmark their performance. TRECVID has been funded by NIST (National Institute of Standards and Technology) and other US government agencies. In addition, many organizations and individuals worldwide contribute significant time and effort. TRECVID 2019 represented a continuation of four tasks from TRECVID 2018. In total, 27 teams from various research organizations worldwide completed one or more of the following four tasks: 1. Ad-hoc Video Search (AVS) 2. Instance Search (INS) 3. Activities in Extended Video (ActEV) 4. Video to Text Description (VTT) This paper is an introduction to the evaluation framework, tasks, data, and measures used in the workshop.

cs.CV↗

arXiv2020

Cautious Adaptation For Reinforcement Learning in Safety-Critical Settings

Reinforcement learning (RL) in real-world safety-critical target settings like urban driving is hazardous, imperiling the RL agent, other agents, and the environment. To overcome this difficulty, we propose a "safety-critical adaptation" task setting: an agent first trains in non-safety-critical "source" environments such as in a simulator, before it adapts to the target environment where failures carry heavy costs. We propose a solution approach, CARL, that builds on the intuition that prior experience in diverse environments equips an agent to estimate risk, which in turn enables relative safety through risk-averse, cautious adaptation. CARL first employs model-based RL to train a probabilistic model to capture uncertainty about transition dynamics and catastrophic states across varied source environments. Then, when exploring a new safety-critical environment with unknown dynamics, the CARL agent plans to avoid actions that could lead to catastrophic states. In experiments on car driving, cartpole balancing, half-cheetah locomotion, and robotic object manipulation, CARL successfully acquires cautious exploration behaviors, yielding higher rewards with fewer failures than strong RL adaptation baselines. Website at https://sites.google.com/berkeley.edu/carl.

cs.LG↗

arXiv2019

Unsupervised Projection Networks for Generative Adversarial Networks

We propose the use of unsupervised learning to train projection networks that project onto the latent space of an already trained generator. We apply our method to a trained StyleGAN, and use our projection network to perform image super-resolution and clustering of images into semantically identifiable groups.

cs.CV↗

arXiv2019

REPLAB: A Reproducible Low-Cost Arm Benchmark Platform for Robotic Learning

Standardized evaluation measures have aided in the progress of machine learning approaches in disciplines such as computer vision and machine translation. In this paper, we make the case that robotic learning would also benefit from benchmarking, and present the "REPLAB" platform for benchmarking vision-based manipulation tasks. REPLAB is a reproducible and self-contained hardware stack (robot arm, camera, and workspace) that costs about 2000 USD, occupies a cuboid of size 70x40x60 cm, and permits full assembly within a few hours. Through this low-cost, compact design, REPLAB aims to drive wide participation by lowering the barrier to entry into robotics and to enable easy scaling to many robots. We envision REPLAB as a framework for reproducible research across manipulation tasks, and as a step in this direction, we define a template for a grasping benchmark consisting of a task definition, evaluation protocol, performance measures, and a dataset of 92k grasp attempts. We implement, evaluate, and analyze several previously proposed grasping approaches to establish baselines for this benchmark. Finally, we also implement and evaluate a deep reinforcement learning approach for 3D reaching tasks on our REPLAB platform. Project page with assembly instructions, code, and videos: https://goo.gl/5F9dP4.

cs.RO↗

arXiv2017

Porcupine Neural Networks: (Almost) All Local Optima are Global

Neural networks have been used prominently in several machine learning and statistics applications. In general, the underlying optimization of neural networks is non-convex which makes their performance analysis challenging. In this paper, we take a novel approach to this problem by asking whether one can constrain neural network weights to make its optimization landscape have good theoretical properties while at the same time, be a good approximation for the unconstrained one. For two-layer neural networks, we provide affirmative answers to these questions by introducing Porcupine Neural Networks (PNNs) whose weight vectors are constrained to lie over a finite set of lines. We show that most local optima of PNN optimizations are global while we have a characterization of regions where bad local optimizers may exist. Moreover, our theoretical and empirical results suggest that an unconstrained neural network can be approximated using a polynomially-large PNN.

stat.ML↗