arXiv subjects

Cheng Yin

Publications and source records attributed to Cheng Yin.

8 recordsLinked to original sources

arXiv2026

SimpleMemVLA: A Simple but Effective Native-Video Memory for Vision-Language-Action Models

Long-horizon manipulation is partially observable: the information needed to choose the next action may appear only in observations from minutes earlier. Existing memory mechanisms: retrieval banks, learned compressors, recurrent states must decide what to keep from the past before knowing what a future decision will require. This was motivated by the assumption that minute-scale history is too large to process directly, which modern VLM backbones no longer make true. In this work, we introduce SimpleMemVLA, a VLA without a dedicated memory module. It keeps the sampled history intact and passes it to the backbone in the timestamped video format the backbone was pretrained to process; the hidden states of a generated sub-task then form the only channel from history to a standard flow-matching action head. Since consecutive decisions share most of their history, prefilling the shared prefix during action execution keeps latency close to a single-frame VLA. SimpleMemVLA sets a new state of the art on four memory benchmarks without cost on general-purpose control. Holding the backbone and training setup fixed, it outperforms retrieval, compression and recurrent-state mechanisms by a wide margin, and causal interventions confirm that the policy genuinely reads its history. Code available at https://github.com/wadeKeith/SimpleMemVLA

cs.CV↗

arXiv2025

DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models

Does Chain-of-Thought (CoT) reasoning genuinely improve Vision Language Action (VLA) models, or does it merely add overhead? Existing CoT-VLA systems report limited and inconsistent gains, yet no prior work has rigorously diagnosed when and why CoT helps robots act. Through systematic experiments, we identify two necessary conditions that must be jointly satisfied for CoT to be effective in VLA: (1) Decoding Alignment: CoT and actions must be generated with modality-appropriate mechanisms; forcing both through a single autoregressive decoder is not merely suboptimal but actively harmful, degrading performance by 4.2 percentage points; (2) Causal Alignment: CoT must be causally linked to task success via outcome-based optimization; without it, supervised CoT is indistinguishable from no reasoning at all under action-execution-sensitive dynamics shift, exhibiting a 32.0 pp performance drop nearly identical to the 31.6 pp drop of a reasoning-free baseline. Guided by these findings, we build DeepThinkVLA: a hybrid-attention decoder satisfies Condition 1 by pairing causal attention for language with bidirectional attention for parallel action decoding, while a two-stage SFT-then-RL pipeline satisfies Condition 2 by aligning the full reasoning: action chain with sparse task-success rewards. DeepThinkVLA achieves 97.0\% success on LIBERO, 79.0\% robustness on LIBERO-Plus (vs. 61.6\% for $\pi_0$-FAST), and 59.3\% success on RoboTwin 2.0, exceeding the strongest baseline by 21.7 points. Furthermore, real-robot experiments provide preliminary evidence for the physical applicability of our CoT data construction and hybrid architecture. Our codes are available at https://github.com/OpenBMB/DeepThinkVLA.

cs.LG↗

arXiv2024

Does Optimal Control Always Benefit from Better Prediction? An Analysis Framework for Predictive Optimal Control

The ``prediction + optimal control'' scheme has shown good performance in many applications of automotive, traffic, robot, and building control. In practice, the prediction results are simply considered correct in the optimal control design process. However, in reality, these predictions may never be perfect. Under a conventional stochastic optimal control formulation, it is difficult to answer questions like ``what if the predictions are wrong''. This paper presents an analysis framework for predictive optimal control where the subjective belief about the future is no longer considered perfect. A novel concept called the hidden prediction state is proposed to establish connections among the predictors, the subjective beliefs, the control policies and the objective control performance. Based on this framework, the predictor evaluation problem is analyzed. Three commonly-used predictor evaluation measures, including the mean squared error, the regret and the log-likelihood, are considered. It is shown that neither using the mean square error nor using the likelihood can guarantee a monotonic relationship between the predictor error and the optimal control cost. To guarantee control cost improvement, it is suggested the predictor should be evaluated with the control performance, e.g., using the optimal control cost or the regret to evaluate predictors. Numerical examples and examples from automotive applications with real-world driving data are provided to illustrate the ideas and the results.

eess.SY↗

arXiv2023

Navigating Open Set Scenarios for Skeleton-based Action Recognition

In real-world scenarios, human actions often fall outside the distribution of training data, making it crucial for models to recognize known actions and reject unknown ones. However, using pure skeleton data in such open-set conditions poses challenges due to the lack of visual background cues and the distinct sparse structure of body pose sequences. In this paper, we tackle the unexplored Open-Set Skeleton-based Action Recognition (OS-SAR) task and formalize the benchmark on three skeleton-based datasets. We assess the performance of seven established open-set approaches on our task and identify their limits and critical generalization issues when dealing with skeleton information. To address these challenges, we propose a distance-based cross-modality ensemble method that leverages the cross-modal alignment of skeleton joints, bones, and velocities to achieve superior open-set recognition performance. We refer to the key idea as CrossMax - an approach that utilizes a novel cross-modality mean max discrepancy suppression mechanism to align latent spaces during training and a cross-modality distance-based logits refinement method during testing. CrossMax outperforms existing approaches and consistently yields state-of-the-art results across all datasets and backbones. The benchmark, code, and models will be released at https://github.com/KPeng9510/OS-SAR.

cs.CV↗

arXiv2023

A Method to Improve the Performance of Reinforcement Learning Based on the Y Operator for a Class of Stochastic Differential Equation-Based Child-Mother Systems

This paper introduces a novel operator, termed the Y operator, to elevate control performance in Actor-Critic(AC) based reinforcement learning for systems governed by stochastic differential equations(SDEs). The Y operator ingeniously integrates the stochasticity of a class of child-mother system into the Critic network's loss function, yielding substantial advancements in the control performance of RL algorithms.Additionally, the Y operator elegantly reformulates the challenge of solving partial differential equations for the state-value function into a parallel problem for the drift and diffusion functions within the system's SDEs.A rigorous mathematical proof confirms the operator's validity.This transformation enables the Y Operator-based Reinforcement Learning(YORL) framework to efficiently tackle optimal control problems in both model-based and data-driven systems.The superiority of YORL is demonstrated through linear and nonlinear numerical examples showing its enhanced performance over existing methods post convergence.

cs.AI↗

arXiv2021

Deep Reinforcement Learning for Intelligent Reflecting Surface-assisted D2D Communications

In this paper, we propose a deep reinforcement learning (DRL) approach for solving the optimisation problem of the network's sum-rate in device-to-device (D2D) communications supported by an intelligent reflecting surface (IRS). The IRS is deployed to mitigate the interference and enhance the signal between the D2D transmitter and the associated D2D receiver. Our objective is to jointly optimise the transmit power at the D2D transmitter and the phase shift matrix at the IRS to maximise the network sum-rate. We formulate a Markov decision process and then propose the proximal policy optimisation for solving the maximisation game. Simulation results show impressive performance in terms of the achievable rate and processing time.

eess.SP↗

arXiv2016

Absence of Anderson localization in certain random lattices

We report on the transition between an Anderson localized regime and a conductive regime in a 1D scattering system with correlated disorder. We show experimentally that when long-range correlations, in the form of a power-law spectral density with power larger than 2, are introduced the localization length becomes much bigger than the sample size and the transmission peaks typical of an Anderson localized system merge into a pass band. As other forms of long-range correlations are known to have the opposite effect, i.e. to enhance localization, our results show that care is needed when discussing the effects of correlations, as different kinds of long-range correlations can give rise to very different behavior.

physics.optics↗

arXiv2016

Pseudopotential MRT lattice Boltzmann model for cavitation bubble collapse with high density ratio

The dynamics of the cavitation bubble collapse is a fundamental issue for the bubble collapse application and prevention. In present work, the modified forcing scheme for the pseudopotential multi-relaxation-time lattice Boltzmann model developed by Li Q. et al. is adopted to develop a cavitation bubble collapse model. In the respects of coexistence curves and Laplace law verification, the improved pseudopotential multi-relaxation-time lattice Boltzmann model is investigated. The independence between the kinematic viscosity and the thermodynamic consistency, surface tension is founded. By homogeneous and heterogeneous cavitation simulation, the capability of the present model to describe the cavitation bubble development as well as the cavitation inception is verified. The bubble collapse between two parallel walls is simulated. The dynamic process of collapsing bubble is consistent with the results from experiments and simulations by other numerical method. It is demonstrated that the present pseudopotential multi-relaxation-time lattice Boltzmann model is available and efficient, and the lattice Boltzmann method is an alternative tool for collapsing bubble modeling.

physics.flu-dyn↗