Yi Ma

Mitsuhiko Nakamoto, Yuexiang Zhai, Anikait Singh et al. · stanford

A compelling use case of offline reinforcement learning (RL) is to obtain a policy initialization from existing datasets followed by fast online fine-tuning with limited interaction. However, existing offline RL methods tend to behave poorly during fine-tuning. In this paper, we devise an approach for learning an effective initialization from offline data that also enables fast online fine-tuning capabilities. Our approach, calibrated Q-learning (Cal-QL), accomplishes this by learning a conservative value function initialization that underestimates the value of the learned policy from offline data, while also being calibrated, in the sense that the learned Q-values are at a reasonable scale. We refer to this property as calibration, and define it formally as providing a lower bound on the true value function of the learned policy and an upper bound on the value of some other (suboptimal) reference policy, which may simply be the behavior policy. We show that offline RL algorithms that learn such calibrated value functions lead to effective online fine-tuning, enabling us to take the benefits of offline initializations in online fine-tuning. In practice, Cal-QL can be implemented on top of the conservative Q learning (CQL) for offline RL within a one-line code change. Empirically, Cal-QL outperforms state-of-the-art methods on 9/11 fine-tuning benchmark tasks that we study in this paper. Code and video are available at https://nakamotoo.github.io/Cal-QL

Haozhi Qi, Ashish Kumar, Roberto Calandra et al. · berkeley

Generalized in-hand manipulation has long been an unsolved challenge of robotics. As a small step towards this grand goal, we demonstrate how to design and learn a simple adaptive controller to achieve in-hand object rotation using only fingertips. The controller is trained entirely in simulation on only cylindrical objects, which then - without any fine-tuning - can be directly deployed to a real robot hand to rotate dozens of objects with diverse sizes, shapes, and weights over the z-axis. This is achieved via rapid online adaptation of the controller to the object properties using only proprioception history. Furthermore, natural and stable finger gaits automatically emerge from training the control policy via reinforcement learning. Code and more videos are available at https://haozhi.io/hora

Kaidi Wang, Yi Ma, Mahdi Boloursaz Mashhadi et al.

This paper studies issues that arise with respect to the joint optimization for convergence time in federated learning over wireless networks (FLOWN). We consider the criterion and protocol for selection of participating devices in FLOWN under the energy constraint and derive its impact on device selection. In order to improve the training efficiency, age-of-information (AoI) enables FLOWN to assess the freshness of gradient updates among participants. Aiming to speed up convergence, we jointly investigate global loss minimization and latency minimization in a Stackelberg game based framework. Specifically, we formulate global loss minimization as a leader-level problem for reducing the number of required rounds, and latency minimization as a follower-level problem to reduce time consumption of each round. By decoupling the follower-level problem into two sub-problems, including resource allocation and sub-channel assignment, we achieve an optimal strategy of the follower through monotonic optimization and matching theory. At the leader-level, we derive an upper bound of convergence rate and subsequently reformulate the global loss minimization problem and propose a new age-of-update (AoU) based device selection algorithm. Simulation results indicate the superior performance of the proposed AoU based device selection scheme in terms of the convergence rate, as well as efficient utilization of available sub-channels.

Druv Pai, Michael Psenka, Chih-Yuan Chiu et al.

We consider the problem of learning discriminative representations for data in a high-dimensional space with distribution supported on or around multiple low-dimensional linear subspaces. That is, we wish to compute a linear injective map of the data such that the features lie on multiple orthogonal subspaces. Instead of treating this learning problem using multiple PCAs, we cast it as a sequential game using the closed-loop transcription (CTRL) framework recently proposed for learning discriminative and generative representations for general low-dimensional submanifolds. We prove that the equilibrium solutions to the game indeed give correct representations. Our approach unifies classical methods of learning subspaces with modern deep learning practice, by showing that subspace learning problems may be provably solved using the modern toolkit of representation learning. In addition, our work provides the first theoretical justification for the CTRL framework, in the important case of linear subspaces. We support our theoretical findings with compelling empirical evidence. We also generalize the sequential game formulation to more general representation learning problems. Our code, including methods for easy reproduction of experimental results, is publically available on GitHub.

Xiaohan Hu, Yi Ma, Chenjun Xiao et al.

One of the fundamental challenges for offline reinforcement learning (RL) is ensuring robustness to data distribution. Whether the data originates from a near-optimal policy or not, we anticipate that an algorithm should demonstrate its ability to learn an effective control policy that seamlessly aligns with the inherent distribution of offline data. Unfortunately, behavior regularization, a simple yet effective offline RL algorithm, tends to struggle in this regard. In this paper, we propose a new algorithm that substantially enhances behavior-regularization based on conservative policy iteration. Our key observation is that by iteratively refining the reference policy used for behavior regularization, conservative policy update guarantees gradually improvement, while also implicitly avoiding querying out-of-sample actions to prevent catastrophic learning failures. We prove that in the tabular setting this algorithm is capable of learning the optimal policy covered by the offline dataset, commonly referred to as the in-sample optimal policy. We then explore several implementation details of the algorithm when function approximations are applied. The resulting algorithm is easy to implement, requiring only a few lines of code modification to existing methods. Experimental results on the D4RL benchmark indicate that our method outperforms previous state-of-the-art baselines in most tasks, clearly demonstrate its superiority over behavior regularization.

Siyan Dong, Zijun Wang, Lulu Cai et al.

Real-time dense scene reconstruction during unstable camera motions is crucial for robotics, yet current RGB-D SLAM systems fail when cameras experience large viewpoint changes, fast motions, or sudden shaking. Classical optimization-based methods deliver high accuracy but fail with poor initialization during large motions, while learning-based approaches provide robustness but lack sufficient accuracy for dense reconstruction. We address this challenge through a combination of learning-based initialization with optimization-based refinement. Our method employs a camera pose regression network to predict metric-aware relative poses from consecutive RGB-D frames, which serve as reliable starting points for a randomized optimization algorithm that further aligns depth images with the scene geometry. Extensive experiments demonstrate promising results: our approach outperforms the best competitor on challenging benchmarks, while maintaining comparable accuracy on stable motion sequences. The system operates in real-time, showcasing that combining simple and principled techniques can achieve both robustness for unstable motions and accuracy for dense reconstruction. Project page: https://github.com/siyandong/PROFusion.

Peiqi Duan, Boyu Li, Yixin Yang et al.

Event cameras are emerging imaging technology that offers advantages over conventional frame-based imaging sensors in dynamic range and sensing speed. Complementing the rich texture and color perception of traditional image frames, the hybrid camera system of event and frame-based cameras enables high-performance imaging. With the assistance of event cameras, high-quality image/video enhancement methods make it possible to break the limits of traditional frame-based cameras, especially exposure time, resolution, dynamic range, and frame rate limits. This paper focuses on five event-aided image and video enhancement tasks (i.e., event-based video reconstruction, event-aided high frame rate video reconstruction, image deblurring, image super-resolution, and high dynamic range image reconstruction), provides an analysis of the effects of different event properties, a real-captured and ground truth labeled benchmark dataset, a unified benchmarking of state-of-the-art methods, and an evaluation for two mainstream event simulators. In detail, this paper collects a real-captured evaluation dataset EventAid for five event-aided image/video enhancement tasks, by using "Event-RGB" multi-camera hybrid system, taking into account scene diversity and spatiotemporal synchronization. We further perform quantitative and visual comparisons for state-of-the-art algorithms, provide a controlled experiment to analyze the performance limit of event-aided image deblurring methods, and discuss open problems to inspire future research.

Ruilong Li, Brent Yi, Junchen Liu et al.

Transformers are increasingly prevalent for multi-view computer vision tasks, where geometric relationships between viewpoints are critical for 3D perception. To leverage these relationships, multi-view transformers must use camera geometry to ground visual tokens in 3D space. In this work, we compare techniques for conditioning transformers on cameras: token-level raymap encodings, attention-level relative pose encodings, and a new relative encoding we propose -- Projective Positional Encoding (PRoPE) -- that captures complete camera frustums, both intrinsics and extrinsics, as a relative positional encoding. Our experiments begin by showing how relative camera conditioning improves performance in feedforward novel view synthesis, with further gains from PRoPE. This holds across settings: scenes with both shared and varying intrinsics, when combining token- and attention-level conditioning, and for generalization to inputs with out-of-distribution sequence lengths and camera intrinsics. We then verify that these benefits persist for different tasks, stereo depth estimation and discriminative spatial cognition, as well as larger model sizes.

Chun-Hsiao Yeh, Yilin Wang, Nanxuan Zhao et al.

Recent diffusion-based image editing methods have significantly advanced text-guided tasks but often struggle to interpret complex, indirect instructions. Moreover, current models frequently suffer from poor identity preservation, unintended edits, or rely heavily on manual masks. To address these challenges, we introduce X-Planner, a Multimodal Large Language Model (MLLM)-based planning system that effectively bridges user intent with editing model capabilities. X-Planner employs chain-of-thought reasoning to systematically decompose complex instructions into simpler, clear sub-instructions. For each sub-instruction, X-Planner automatically generates precise edit types and segmentation masks, eliminating manual intervention and ensuring localized, identity-preserving edits. Additionally, we propose a novel automated pipeline for generating large-scale data to train X-Planner which achieves state-of-the-art results on both existing benchmarks and our newly introduced complex editing benchmark.

Chris Junchi Li, Yaodong Yu, Nicolas Loizou et al.

We study the stochastic bilinear minimax optimization problem, presenting an analysis of the same-sample Stochastic ExtraGradient (SEG) method with constant step size, and presenting variations of the method that yield favorable convergence. In sharp contrasts with the basic SEG method whose last iterate only contracts to a fixed neighborhood of the Nash equilibrium, SEG augmented with iteration averaging provably converges to the Nash equilibrium under the same standard settings, and such a rate is further improved by incorporating a scheduled restarting procedure. In the interpolation setting where noise vanishes at the Nash equilibrium, we achieve an optimal convergence rate up to tight constants. We present numerical experiments that validate our theoretical findings and demonstrate the effectiveness of the SEG method when equipped with iteration averaging and restarting.

Chaobing Song, Yong Jiang, Yi Ma

In this paper, through a very intuitive vanilla proximal method perspective, we derive accelerated high-order optimization algorithms for minimizing a convex function that has Hölder continuous derivatives. In this general convex setting, we propose a concise unified acceleration framework (UAF), which reconciles the two different high-order acceleration approaches, one by Nesterov and Baes [29, 3, 33] and one by Monteiro and Svaiter [25]. As result, the UAF unifies the high-order acceleration instances [29, 3, 33, 15, 16, 25, 19, 6, 14] of the two approaches by only two problem-related parameters and two additional parameters for framework design. Furthermore, the UAF (and its analysis) is the first approach to make high-order methods applicable for high-order smoothness conditions with respect to non-Euclidean norms. If the function is further uniformly convex, we propose a general restart scheme for the UAF. The iteration complexities of instances of both the UAF and the restarted UAF match existing lower bounds in most important cases [2, 16]. For practical implementation, we introduce a new and effective heuristic that significantly simplifies the binary search procedure required by the framework. We use experiments to verify the effectiveness of the heuristic and demonstrate clear and consistent advantages of high-order acceleration methods over first-order ones, in terms of run-time complexity. Finally, the UAF is proposed directly in the general composite convex setting, thus show that the existing high-order algorithms [29, 3, 33, 16, 6, 14] can be naturally extended to the general composite convex setting.