More Like this

1. Learning by Cheating

   Dian Chen, Brady Zhou ( October 2019 , Conference on Robot Learning)

   Vision-based urban driving is hard. The autonomous system needs to learn to perceive the world and act in it. We show that this challenging learning problem can be simplified by decomposing it into two stages. We first train an agent that has access to privileged information. This privileged agent cheats by observing the ground-truth layout of the environment and the positions of all traffic participants. In the second stage, the privileged agent acts as a teacher that trains a purely vision-based sensorimotor agent. The resulting sensorimotor agent does not have access to any privileged information and does not cheat. This two-stage training procedure is counter-intuitive at first, but has a number of important advantages that we analyze and empirically demonstrate. We use the presented approach to train a vision-based autonomous driving system that substantially outperforms the state of the art on the CARLA benchmark and the recent NoCrash benchmark. Our approach achieves, for the first time, 100% success rate on all tasks in the original CARLA benchmark, sets a new record on the NoCrash benchmark, and reduces the frequency of infractions by an order of magnitude compared to the prior state of the art. 

   more » « less
2. A correct-and-certify approach to self-supervise object pose estimators via ensemble self-training

   J. Shil ; R. Talak ; D. Maggio ; L. Carlone ( July 2023 , Robotics: Science and Systems (RSS))

   Real-world robotics applications demand object pose estimation methods that work reliably across a variety of scenarios. Modern learning-based approaches require large labeled datasets and tend to perform poorly outside the training domain. Our first contribution is to develop a robust corrector module that corrects pose estimates using depth information, thus enabling existing methods to better generalize to new test domains; the corrector operates on semantic keypoints (but is also applicable to other pose estimators) and is fully differentiable. Our second contribution is an ensemble self-training approach that simultaneously trains multiple pose estimators in a self-supervised manner. Our ensemble self-training architecture uses the robust corrector to refine the output of each pose estimator; then, it evaluates the quality of the outputs using observable correctness certificates; finally, it uses the observably correct outputs for further training, without requiring external supervision. As an additional contribution, we propose small improvements to a regression-based keypoint detection architecture, to enhance its robustness to outliers; these improvements include a robust pooling scheme and a robust centroid computation. Experiments on the YCBV and TLESS datasets show the proposed ensemble self-training outperforms fully supervised baselines while not requiring 3D annotations on real data. 

   more » « less
3. Curriculum Based Reinforcement Learning of Grid Topology Controllers to Prevent Thermal Cascading

   https://doi.org/10.1109/TPWRS.2022.3213487  

   Matavalam, Amarsagar Reddy ; Guddanti, Kishan Prudhvi ; Weng, Yang ; Ajjarapu, Venkataramana ( October 2022 , IEEE Transactions on Power Systems)

   This paper describes how domain knowledge of power system operators can be integrated into reinforcement learning (RL) frameworks to effectively learn agents that control the grid's topology to prevent thermal cascading. Typical RL-based topology controllers fail to perform well due to the large search/optimization space. Here, we propose an actor-critic-based agent to address the problem's combinatorial nature and train the agent using the RL environment developed by RTE, the French TSO. To address the challenge of the large optimization space, a curriculum-based approach with reward tuning is incorporated into the training procedure by modifying the environment using network physics for enhanced agent learning. Further, a parallel training approach on multiple scenarios is employed to avoid biasing the agent to a few scenarios and make it robust to the natural variability in grid operations. Without these modifications to the training procedure, the RL agent failed for most test scenarios, illustrating the importance of properly integrating domain knowledge of physical systems for real-world RL learning. The agent was tested by RTE for the 2019 learning to run the power network challenge and was awarded the 2nd place in accuracy and 1st place in speed. The developed code is open-sourced for public use. Analysis of a simple system proves the enhancement in training RL-agents using the curriculum. 

   more » « less
4. CYCLE: Learning to Self-Refine the Code Generation

   https://doi.org/10.1145/3649825  

   Ding, Yangruibo ; Min, Marcus J ; Kaiser, Gail ; Ray, Baishakhi ( April 2024 , Proceedings of the ACM on Programming Languages)

   Pre-trained code language models have achieved promising performance in code generation and improved the programming efficiency of human developers. However, their self-refinement capability is typically overlooked by the existing evaluations of code LMs, which focus only on the accuracy of the one-time prediction. For the cases when code LMs fail to implement the correct program, developers actually find it hard to debug and fix the faulty prediction since it is not written by the developers themselves. Unfortunately, our study reveals that code LMs cannot efficiently self-refine their faulty generations as well.

   In this paper, we propose CYCLE framework, learning to self-refine the faulty generation according to the available feedback, such as the execution results reported by the test suites. We evaluate CYCLE on three popular code generation benchmarks, HumanEval, MBPP, and APPS. The results reveal that CYCLE successfully maintains, sometimes improves, the quality of one-time code generation, while significantly improving the self-refinement capability of code LMs. We implement four variants of CYCLE with varied numbers of parameters across 350M, 1B, 2B, and 3B, and the experiments show that CYCLE consistently boosts the code generation performance, by up to 63.5 

   more » « less
5. Learning to Optimize Differentiable Games

   Chen, Xuxi ; Vadori, Nelson ; Chen, Tianlong ; Wang, Zhangyang ( July 2023 , International Conference on Machine Learning)

   Many machine learning problems can be abstracted in solving game theory formulations and boil down to optimizing nested objectives, such as generative adversarial networks (GANs) and multi-agent reinforcement learning. Solving these games requires finding their stable fixed points or Nash equilibrium. However, existing algorithms for solving games suffer from empirical instability, hence demanding heavy ad-hoc tuning in practice. To tackle these challenges, we resort to the emerging scheme of Learning to Optimize (L2O), which discovers problem-specific efficient optimization algorithms through data-driven training. Our customized L2O framework for differentiable game theory problems, dubbed “Learning to Play Games" (L2PG), seeks a stable fixed point solution, by predicting the fast update direction from the past trajectory, with a novel gradient stability-aware, sign-based loss function. We further incorporate curriculum learning and self-learning to strengthen the empirical training stability and generalization of L2PG. On test problems including quadratic games and GANs, L2PG can substantially accelerate the convergence, and demonstrates a remarkably more stable trajectory. Codes are available at https://github.com/VITA-Group/L2PG. 

   more » « less