arXiv subjects

Yicheng Zeng

Publications and source records attributed to Yicheng Zeng.

16 recordsLinked to original sources

arXiv2026

ADMM-Based Safety-Critical Distributed NMPC for Cooperative Transportation by Quadrupedal Robots

This paper presents a safety-critical distributed nonlinear model predictive control (DNMPC) framework for cooperative payload transportation by teams of quadrupedal robots. The proposed approach models the robotic team and the shared payload as a dynamically coupled networked system with rigid holonomic coupling constraints arising from cooperative transportation. To enable distributed real-time optimization, the centralized finite-horizon optimal control problem is decomposed into parallel local NMPC subproblems coordinated through the alternating direction method of multipliers (ADMM). The resulting distributed framework enforces consensus over both payload-state and interaction-wrench trajectories while explicitly incorporating acceleration-level holonomic coupling constraints within the distributed predictive control formulation. Safety-critical obstacle avoidance constraints for both the robotic agents and payload are enforced using higher-order control barrier functions (HOCBFs). The framework is validated through numerical simulations with teams of two, three, and four quadrupedal robots transporting shared payloads in cluttered environments. Real-time experiments on two- and three-robot teams demonstrate safe and robust transportation under payload uncertainty and external disturbances. Compared with centralized NMPC, the proposed framework achieves up to 23% reduction in average NLP solve time while maintaining comparable closed-loop performance. Ablation studies further demonstrate robustness to communication delays and show that explicit payload-state consensus and holonomic constraints substantially improve payload tracking and distributed coordination over existing wrench-only consensus formulations.

cs.RO↗

arXiv2026

A Unified Framework for In-Context Learning with Causal and Masked Language Models

In-context learning (ICL) has emerged as a central capability of pretrained language models, yet its theoretical analysis has focused primarily on causal language models trained by left-to-right autoregressive prediction, such as GPT-style models. Masked language models instead recover masked tokens from bidirectional context, and their role in ICL remains less understood. We develop a statistical learning framework that represents the context examples by their empirical measure and models prediction as a function of the context and the query. This formulation places autoregressive and masked pretraining objectives within a common excess-risk analysis. Under Wasserstein-type regularity conditions, we relate pretraining with T tasks and N samples per task to k-shot excess risk at inference, obtaining same-order upper bounds for masked and autoregressive objectives. We also study task-distribution shift, where pretraining tasks are sampled from P and inference tasks from Q; the resulting bound contains an additional term controlled by the lifted Wasserstein distance between P and Q. The bounds further imply an order-optimal allocation under a fixed pretraining data budget and refined rates under intrinsic low-dimensional structure. Experiments on controlled function-learning tasks show that the Masked Pair Encoder (MPE) can achieve performance comparable to GPT-2-style causal Transformers, suggesting that ICL behavior is not specific to causal language models.

cs.LG↗

arXiv2026

Binary Decompilation LLM with Feedback-Driven Multi-Turn Refinement

Binary decompilation is fundamental to security tasks such as vulnerability discovery, malware inspection, and executable-only program understanding. Recent LLM-based decompilation methods have shown promising results, but most still follow a single-turn generation paradigm: given assembly code or decompiler-produced pseudo-code, the model generates one output and stops. Consequently, the generated code may appear readable or even compile successfully, yet still deviate from the behavior of the original binary and mislead downstream analysis. This paper presents AutoDecompiler, a decompilation-specialized LLM trained with reinforcement learning for feedback-driven multi-turn binary decompilation. Instead of treating decompilation as one-shot code generation, AutoDecompiler formulates it as an iterative refinement process, where the model revises generated code based on compilation, execution, and input/output testing feedback. To enable this process, we design decompilation-specific rewards that capture code validity, recompilability, execution consistency, and semantic fidelity. We further construct stage-aware diagnostic feedback from compiler errors, execution failures, and failed test cases, and introduce progress-aware trajectory rewarding and turn-aware advantage reweighting to encourage beneficial revisions while suppressing regressions. We train the AutoDecompiler family and evaluate it across different input settings, model scales, and benchmarks. Experimental results show that AutoDecompiler consistently outperforms its single-turn counterparts under the same model size and input setting, achieving clear improvements in behavioral re-executability. These results demonstrate that learning to exploit program feedback with reinforcement learning is an effective direction for improving the functional correctness of LLM-based binary decompilation.

cs.SE↗

arXiv2026

Missingness-Adaptive Factor Identification in High-Dimensional Data

Determining the number of factors in high-dimensional factor models remains a fundamental challenge, particularly when data are incomplete. This paper introduces the concept of identifiable factors, those that can be reliably recovered despite missing observations, and proposes the Missingness-Adaptive Thresholding Estimator (MATE). To our knowledge, MATE is the first missingness-adaptive framework for factor number determination that accommodates both homogeneous and heterogeneous missingness without imposing restrictive assumptions on factor strength. Notably, it operates without data imputation, circumventing the computational burden associated with most existing approaches. We establish a rigorous theoretical foundation for MATE, proving its consistency under a range of structural conditions. Extensive simulations and real-world applications demonstrate that MATE consistently outperforms state-of-the-art methods, exhibiting superior robustness in settings with high missingness rates and weak factor signals.

stat.ME↗

arXiv2026

Safety-Critical Centralized Nonlinear MPC for Cooperative Payload Transportation by Two Quadrupedal Robots

This paper presents a safety-critical centralized nonlinear model predictive control (NMPC) framework for cooperative payload transportation by two quadrupedal robots. The interconnected robot-payload system is modeled as a discrete-time nonlinear differential-algebraic system, capturing the coupled dynamics through holonomic constraints and interaction wrenches. To ensure safety in complex environments, we develop a control barrier function (CBF)-based NMPC formulation that enforces collision avoidance constraints for both the robots and the payload. The proposed approach retains the interaction wrenches as decision variables, resulting in a structured DAE-constrained optimal control problem that enables efficient real-time implementation. The effectiveness of the algorithm is validated through extensive hardware experiments on two Unitree Go2 platforms performing cooperative payload transportation in cluttered environments under mass and inertia uncertainty and external push disturbances.

cs.RO↗

arXiv2026

ADMM-Based Distributed MPC with Control Barrier Functions for Safe Multi-Robot Quadrupedal Locomotion

This paper proposes a fully decentralized model predictive control (MPC) framework with control barrier function (CBF) constraints for safety-critical trajectory planning in multi-robot legged systems. The incorporation of CBF constraints introduces explicit inter-agent coupling, which prevents direct decomposition of the resulting optimal control problems. To address this challenge, we reformulate the centralized safety-critical MPC problem using a structured distributed optimization framework based on the alternating direction method of multipliers (ADMM). By introducing a novel node-edge splitting formulation with consensus constraints, the proposed approach decomposes the global problem into independent node-local and edge-local quadratic programs that can be solved in parallel using only neighbor-to-neighbor communication. This enables fully decentralized trajectory optimization with symmetric computational load across agents while preserving safety and dynamic feasibility. The proposed framework is integrated into a hierarchical locomotion control architecture for quadrupedal robots, combining high-level distributed trajectory planning, mid-level nonlinear MPC enforcing single rigid body dynamics, and low-level whole-body control enforcing full-order robot dynamics. The effectiveness of the proposed approach is demonstrated through hardware experiments on two Unitree Go2 quadrupedal robots and numerical simulations involving up to four robots navigating uncertain environments with rough terrain and external disturbances. The results show that the proposed distributed formulation achieves performance comparable to centralized MPC while reducing the average per-cycle planning time by up to 51% in the four-agent case, enabling efficient real-time decentralized implementation.

cs.RO↗

arXiv2025

In-Context Learning as Nonparametric Conditional Probability Estimation: Risk Bounds and Optimality

This paper investigates the expected excess risk of in-context learning (ICL) for multiclass classification. We formalize each task as a sequence of labeled examples followed by a query input; a pretrained model then estimates the query's conditional class probabilities. The expected excess risk is defined as the average truncated Kullback-Leibler (KL) divergence between the predicted and true conditional class distributions over a specified family of tasks. We establish a new oracle inequality for this risk, based on KL divergence, in multiclass classification. This yields tight upper and lower bounds for transformer-based models, showing that the ICL estimator achieves the minimax optimal rate (up to logarithmic factors) for conditional probability estimation. From a technical standpoint, our results introduce a novel method for controlling generalization error via uniform empirical entropy. We further demonstrate that multilayer perceptrons (MLPs) can also perform ICL and attain the same optimal rate (up to logarithmic factors) under suitable assumptions, suggesting that effective ICL need not be exclusive to transformer architectures.

stat.ML↗

arXiv2025

Testing for large-dimensional covariance matrix under differential privacy

The increasing prevalence of high-dimensional data across various applications has raised significant privacy concerns in statistical inference. In this paper, we propose a differentially private integrated statistic for testing large-dimensional covariance structures, enabling accurate statistical insights while safeguarding privacy. First, we analyze the global sensitivity of sample eigenvalues for sub-Gaussian populations, where our method bypasses the commonly assumed boundedness of data covariates. For sufficiently large sample size, the privatized statistic guarantees privacy with high probability. Furthermore, when the ratio of dimension to sample size, $d/n \to y \in (0, \infty)$, the privatized test is asymptotically distribution-free with well-known critical values, and detects the local alternative hypotheses distinct from the null at the fastest rate of $1/\sqrt{n}$. Extensive numerical studies on synthetic and real data showcase the validity and powerfulness of our proposed method.

stat.ME↗

arXiv2024

Reference-Steering via Data-Driven Predictive Control for Hyper-Accurate Robotic Flying-Hopping Locomotion

State-of-the-art model-based control designs have been shown to be successful in realizing dynamic locomotion behaviors for robotic systems. The precision of the realized behaviors in terms of locomotion performance via fly, hopping, or walking has not yet been well investigated, despite the fact that the difference between the robot model and physical hardware is doomed to produce inaccurate trajectory tracking. To address this inaccuracy, we propose a referencing-steering method to bridge the model-to-real gap by establishing a data-driven input-output (DD-IO) model on top of the existing model-based design. The DD-IO model takes the reference tracking trajectories as the input and the realized tracking trajectory as the output. By utilizing data-driven predictive control, we steer the reference input trajectories online so that the realized output ones match the actual desired ones. We demonstrate our method on the robot PogoX to realize hyper-accurate hopping and flying behaviors in both simulation and hardware. This data-driven reference-steering approach is straightforward to apply to general robotic systems for performance improvement via hyper-accurate trajectory tracking.

cs.RO↗

arXiv2024

STRIDE: An Open-Source, Low-Cost, and Versatile Bipedal Robot Platform for Research and Education

In this paper, we present STRIDE, a Simple, Terrestrial, Reconfigurable, Intelligent, Dynamic, and Educational bipedal platform. STRIDE aims to propel bipedal robotics research and education by providing a cost-effective implementation with step-by-step instructions for building a bipedal robotic platform while providing flexible customizations via a modular and durable design. Moreover, a versatile terrain setup and a quantitative disturbance injection system are augmented to the robot platform to replicate natural terrains and push forces that can be used to evaluate legged locomotion in practical and adversarial scenarios. We demonstrate the functionalities of this platform by realizing an adaptive step-to-step dynamics based walking controller to achieve dynamic walking. Our work with the open-soured implementation shows that STRIDE is a highly versatile and durable platform that can be used in research and education to evaluate locomotion algorithms, mechanical designs, and robust and adaptative controls.

cs.RO↗

arXiv2024

Enhancing Out-of-Distribution Detection with Multitesting-based Layer-wise Feature Fusion

Deploying machine learning in open environments presents the challenge of encountering diverse test inputs that differ significantly from the training data. These out-of-distribution samples may exhibit shifts in local or global features compared to the training distribution. The machine learning (ML) community has responded with a number of methods aimed at distinguishing anomalous inputs from original training data. However, the majority of previous studies have primarily focused on the output layer or penultimate layer of pre-trained deep neural networks. In this paper, we propose a novel framework, Multitesting-based Layer-wise Out-of-Distribution (OOD) Detection (MLOD), to identify distributional shifts in test samples at different levels of features through rigorous multiple testing procedure. Our approach distinguishes itself from existing methods as it does not require modifying the structure or fine-tuning of the pre-trained classifier. Through extensive experiments, we demonstrate that our proposed framework can seamlessly integrate with any existing distance-based inspection method while efficiently utilizing feature extractors of varying depths. Our scheme effectively enhances the performance of out-of-distribution detection when compared to baseline methods. In particular, MLOD-Fisher achieves superior performance in general. When trained using KNN on CIFAR10, MLOD-Fisher significantly lowers the false positive rate (FPR) from 24.09% to 7.47% on average compared to merely utilizing the features of the last layer.

cs.LG↗

arXiv2024

Adjusting Dynamics of Hopfield Neural Network via Time-variant Stimulus

As a paradigmatic model for nonlinear dynamics studies, the Hopfield Neural Network (HNN) demonstrates a high susceptibility to external disturbances owing to its intricate structure. This paper delves into the challenge of modulating HNN dynamics through time-variant stimuli. The effects of adjustments using two distinct types of time-variant stimuli, namely the Weight Matrix Stimulus (WMS) and the State Variable Stimulus (SVS), along with a Constant Stimulus (CS) are reported. The findings reveal that deploying four WMSs enables the HNN to generate either a four-scroll or a coexisting two-scroll attractor. When combined with one SVS, four WMSs can lead to the formation of an eight-scroll or four-scroll attractor, while the integration of four WMSs and multiple SVSs can induce grid-multi-scroll attractors. Moreover, the introduction of a CS and an SVS can significantly disrupt the dynamic behavior of the HNN. Consequently, suitable adjustment methods are crucial for enhancing the network's dynamics, whereas inappropriate applications can lead to the loss of its chaotic characteristics. To empirically validate these enhancement effects, the study employs an FPGA hardware platform. Subsequently, an image encryption scheme is designed to demonstrate the practical application benefits of the dynamically adjusted HNN in secure multimedia communication. This exploration into the dynamic modulation of HNN via time-variant stimuli offers insightful contributions to the advancement of secure communication technologies.

cs.NE↗

arXiv2023

Sketched Ridgeless Linear Regression: The Role of Downsampling

Overparametrization often helps improve the generalization performance. This paper presents a dual view of overparametrization suggesting that downsampling may also help generalize. Focusing on the proportional regime $m\asymp n \asymp p$, where $m$ represents the sketching size, $n$ is the sample size, and $p$ is the feature dimensionality, we investigate two out-of-sample prediction risks of the sketched ridgeless least square estimator. Our findings challenge conventional beliefs by showing that downsampling does not always harm generalization but can actually improve it in certain cases. We identify the optimal sketching size that minimizes out-of-sample prediction risks and demonstrate that the optimally sketched estimator exhibits stabler risk curves, eliminating the peaks of those for the full-sample estimator. To facilitate practical implementation, we propose an empirical procedure to determine the optimal sketching size. Finally, we extend our analysis to cover central limit theorems and misspecified models. Numerical studies strongly support our theory.

math.ST↗

arXiv2021

Asteria: Deep Learning-based AST-Encoding for Cross-platform Binary Code Similarity Detection

Binary code similarity detection is a fundamental technique for many security applications such as vulnerability search, patch analysis, and malware detection. There is an increasing need to detect similar code for vulnerability search across architectures with the increase of critical vulnerabilities in IoT devices. The variety of IoT hardware architectures and software platforms requires to capture semantic equivalence of code fragments in the similarity detection. However, existing approaches are insufficient in capturing the semantic similarity. We notice that the abstract syntax tree (AST) of a function contains rich semantic information. Inspired by successful applications of natural language processing technologies in sentence semantic understanding, we propose a deep learning-based AST-encoding method, named ASTERIA, to measure the semantic equivalence of functions in different platforms. Our method leverages the Tree-LSTM network to learn the semantic representation of a function from its AST. Then the similarity detection can be conducted efficiently and accurately by measuring the similarity between two representation vectors. We have implemented an open-source prototype of ASTERIA. The Tree-LSTM model is trained on a dataset with 1,022,616 function pairs and evaluated on a dataset with 95,078 function pairs. Evaluation results show that our method outperforms the AST-based tool Diaphora and the-state-of-art method Gemini by large margins with respect to the binary similarity detection. And our method is several orders of magnitude faster than Diaphora and Gemini for the similarity calculation. In the application of vulnerability search, our tool successfully identified 75 vulnerable functions in 5,979 IoT firmware images.

cs.CR↗

arXiv2020

Limiting laws for extreme eigenvalues of large-dimensional spiked Fisher matrices with a divergent number of spikes

Consider the $p\times p$ matrix that is the product of a population covariance matrix and the inverse of another population covariance matrix. Suppose that their difference has a divergent rank with respect to $p$, when two samples of sizes $n$ and $T$ from the two populations are available, we construct its corresponding sample version. In the regime of high dimension where both $n$ and $T$ are proportional to $p$, we investigate the limiting laws for extreme (spiked) eigenvalues of the sample (spiked) Fisher matrix when the number of spikes is divergent and these spikes are unbounded.

math.ST↗

arXiv2019

Order Determination for Spiked Models

Motivated by dimension reduction in regression analysis and signal detection, we investigate the order determination for large dimension matrices including spiked models of which the numbers of covariates are proportional to the sample sizes for different models. Because the asymptotic behaviour of the estimated eigenvalues of the corresponding matrices differ completely from those in fixed dimension scenarios, we then discuss the largest possible number we can identify and introduce a "valley-cliff" criterion. We propose two versions of the criterion: one based on the original differences of eigenvalues and the other based on the transformed differences, which reduces the effects of ridge selection in the former one. This generic method is very easy to implement and computationally inexpensive, and it can be applied to various matrices. As examples, we focus on spiked population models, spiked Fisher matrices and factor models with auto-covariance matrices. Numerical studies are conducted to examine the method's finite sample performances and to compare it with existing methods.

stat.ME↗