arXiv subjects

Yicheng Chen

Publications and source records attributed to Yicheng Chen.

At least 19 recordsLinked to original sources

arXiv2026

Procedural Graphs: Self-Evolving Execution Structures for LLM Agents

Large language models are increasingly deployed as agents that plan over long horizons and act through external tools. Most agents select actions through unconstrained generation over an accumulating history, leaving implicit the procedural knowledge of what to do, in what order, and under which conditions. As trajectories lengthen, agents can lose track of their objectives, invoke tools out of order, and repeat unproductive actions. We introduce the Procedural Graph: just as a knowledge graph organizes factual knowledge into (entity, relation, entity) triplets for what-is questions, a Procedural Graph organizes procedural knowledge into (procedure, relation, procedure) triplets for what-to-do questions. At each decision step, the framework localizes the agent's active node, and a guidance model translates the surrounding subgraph into step-level situational guidance that biases the solver's next action without dictating it. The graph is self-evolving: an LLM refiner contrasts failed trajectories with successful ones and edits the graph's topology and attributes, committing edits that preserve or improve held-out validation performance while retaining rejected ones to discourage repetition. Starting from a minimal skeleton, the loop builds graphs that match or surpass hand-designed ones. It can also repair a flawed expert prior. Across multiple datasets, task types, and LLMs, the Procedural Graph delivers consistent gains over memory-based baselines, and self-evolution further improves performance without manual engineering.

cs.AI↗

arXiv2026

Effector-Centric NMPC of Tiltable-Multirotors for Offset-Free Omnidirectional Aerial Manipulation

Aerial manipulation extends robotic operations to previously inaccessible aerial environments. Unlike arm-equipped aerial systems, tiltable-multirotors can directly generate six-degree-of-freedom wrenches through their flight bases, enabling both efficient movement and omnidirectional operation by tilting the thrust direction. This work presents a design analysis and a wrench-based control framework for tiltable-multirotors in aerial manipulation. We show that a four-rotor tiltable configuration provides a balance between interference-free propeller sizing and hovering efficiency across different attitudes, and its null-space redundancy is crucial for traversing singular configurations under physical constraints. We further show that an upward end-effector placement yields a favorable trade-off between geometric clearance and available wrench. To address disturbances, we propose a dual strategy consisting of a modified integral term for model error and an acceleration-based estimator for external wrenches. Building on these insights, we develop an effector-centric nonlinear model predictive control (NMPC) framework that integrates design choices, singularity handling, and disturbance compensation into a unified formulation. The proposed framework runs fully onboard at 100 Hz on a custom-built tiltable-quadrotor. Real-world experiments, including a 90-deg step cartwheel rotation, whiteboard pushing, and continuous 360-deg valve turning, demonstrate the feasibility of wrench-based omnidirectional manipulation with singularity traversal on a one-DoF-per-arm tiltable-quadrotor.

cs.RO↗

arXiv2026

Hybrid Impedance-Admittance Control with Multi-Link Aerial Robot for Contact-Rich Surface Sliding Task

Multi-link aerial robots can actively deform their articulated structures during flight, giving them strong potential for aerial manipulation. However, they still face substantial challenges in contact-rich aerial manipulation tasks such as surface sliding, which requires both disturbance robustness and compliance to uncertain surface geometry. Force-control strategies such as impedance and admittance control are commonly employed to address these requirements. Although impedance control can provide disturbance-resistant interaction and admittance control can offer compliant adaptation, their opposite force--motion causalities prevent their simultaneous implementation when applied through the same actuation source, such as the rotor thrusts used by conventional aerial robots. To overcome this limitation, we propose a hybrid impedance--admittance control strategy for a multi-link aerial robot. The articulated morphology enables a functional separation of force and motion regulation across joint and rotor actuation sources. In this framework, admittance behavior is generated through joint angle regulation to enhance adaptive interaction, while impedance behavior is achieved by modulating rotor thrust to regulate the sliding motion. This structural coordination allows the robot to leverage the complementary strengths of both control paradigms. As a result, the multi-link aerial robot achieves resilient and adaptive surface sliding. Experimental results demonstrate robust and compliant sliding performance on unknown surfaces.

cs.RO↗

arXiv2026

CoFL-S: Spatially Queryable Sector Flow Fields for Local Language-Conditioned Navigation

Vision-Language Navigation has increasingly emphasized high-level instruction reasoning, memory, global map construction, and instruction decomposition, while the low-level action representation remains comparatively underexplored. We propose CoFL-S, a low-level vision-language-action framework that predicts a language-conditioned flow field over the robot's local visible sector and generates continuous trajectories by rolling out the predicted field. To train this low-level representation, we convert each VLN-CE episode, originally a whole-episode instruction paired with an action sequence, into frame-level local supervision with aligned sub-instructions and matched action, trajectory, and dense flow-field targets. For evaluation, we introduce a continuous-time Habitat benchmark that isolates low-level action interfaces from instruction decomposition and executes all methods through a shared velocity-command controller, enabling decomposition-independent closed-loop comparison across different planner frequencies rather than fixed discrete forward-and-turn transitions in VLN-CE. Under matched encoders and training settings, CoFL-S consistently outperforms action-token and action-chunk baselines across planner frequencies in the continuous-time Habitat benchmark, and zero-shot real-world closed-loop deployment further shows its advantage over both baselines beyond simulation.

cs.RO↗

arXiv2026

MatterSim-MT: A multi-task foundation model for in silico materials characterization

Accurate property characterization is a major bottleneck in materials design. While first-principles methods and task-specific machine-learning models have driven important progress, they remain fundamentally limited in scalability and generalizability across the vast space of structures and properties relevant to real-world materials design. We present MatterSim-MT, a multi-task foundation model for in silico materials simulation and property characterization. The model is pretrained on over 35 million first-principles-labeled structures covering 89 elements, temperatures up to 5000 K and pressures up to 1000 GPa, and is fine-tuned on various properties including Bader charges, magnetic moments, Born effective charges, and dielectric matrices. Out of the box, MatterSim-MT not only serves as a foundation model for predicting material structure, dynamics and thermodynamics, its multi-task architecture also enables a wide range of complex simulations that cannot be captured by potential energy surfaces alone. For example, we demonstrate pressure-dependent LO-TO phonon splitting in SiC with close agreement with experiment, electric hysteresis in ferroelectric BaTiO3, and the cationic-to-anionic redox transition during delithiation of a Li-rich cathode material. Finally, we show that MatterSim-MT scales well with more data and parameters, can be efficiently fine-tuned to higher levels of theory, and can be efficiently extended to new systems via active learning. Overall, we believe this approach provides a scalable route to accurate in silico materials characterization.

cond-mat.mtrl-sci↗

arXiv2026

TREX: Automating LLM Fine-tuning via Agent-Driven Tree-based Exploration

While Large Language Models (LLMs) have empowered AI research agents to perform isolated scientific tasks, automating complex, real-world workflows, such as LLM training, remains a significant challenge. In this paper, we introduce TREX, a multi-agent system that automates the entire LLM training life-cycle. By orchestrating collaboration between two core modules-the Researcher and the Executor-the system seamlessly performs requirement analysis, open-domain literature and data research, formulation of training strategies, preparation of data recipes, and model training and evaluation. The multi-round experimental process is modeled as a search tree, enabling the system to efficiently plan exploration paths, reuse historical results, and distill high-level insights from iterative trials. To evaluate the capability of automated LLM training, we construct FT-Bench, a benchmark comprising 10 tasks derived from real-world scenarios, ranging from optimizing fundamental model capabilities to enhancing performance on domain-specific tasks. Experimental results demonstrate that the TREX agent consistently optimizes model performance on target tasks.

cs.AI↗

arXiv2026

Kernel-Smith: A Unified Recipe for Evolutionary Kernel Optimization

We present Kernel-Smith, a framework for high-performance GPU kernel and operator generation that combines a stable evaluation-driven evolutionary agent with an evolution-oriented post-training recipe. On the agent side, Kernel-Smith maintains a population of executable candidates and iteratively improves them using an archive of top-performing and diverse programs together with structured execution feedback on compilation, correctness, and speedup. To make this search reliable, we build backend-specific evaluation services for Triton on NVIDIA GPUs and Maca on MetaX GPUs. On the training side, we convert long-horizon evolution trajectories into step-centric supervision and reinforcement learning signals by retaining correctness-preserving, high-gain revisions, so that the model is optimized as a strong local improver inside the evolutionary loop rather than as a one-shot generator. Under a unified evolutionary protocol, Kernel-Smith-235B-RL achieves state-of-the-art overall performance on KernelBench with Nvidia Triton backend, attaining the best average speedup ratio and outperforming frontier proprietary models including Gemini-3.0-pro and Claude-4.6-opus. We further validate the framework on the MetaX MACA backend, where our Kernel-Smith-MACA-30B surpasses large-scale counterparts such as DeepSeek-V3.2-think and Qwen3-235B-2507-think, highlighting potential for seamless adaptation across heterogeneous platforms. Beyond benchmark results, the same workflow produces upstream contributions to production systems including SGLang and LMDeploy, demonstrating that LLM-driven kernel optimization can transfer from controlled evaluation to practical deployment.

cs.CL↗

arXiv2026

Design, Modeling and Direction Control of a Wire-Driven Robotic Fish Based on a 2-DoF Crank-Slider Mechanism

Robotic fish have attracted growing attention in recent years owing to their biomimetic design and potential applications in environmental monitoring and biological surveys. Among robotic fish employing the Body-Caudal Fin (BCF) locomotion pattern, motor-driven actuation is widely adopted. Some approaches utilize multiple servo motors to achieve precise body curvature control, while others employ a brushless motor to drive the tail via wire or rod, enabling higher oscillation and swimming speeds. However, the former approaches typically result in limited swimming speed, whereas the latter suffer from poor maneuverability, with few capable of smooth turning. To address this trade-off, we develop a wire-driven robotic fish equipped with a 2-degree-of-freedom (DoF) crank-slider mechanism that decouples propulsion from steering, enabling both high swimming speed and agile maneuvering. In this paper, we first present the design of the robotic fish, including the elastic skeleton, waterproof structure, and the actuation mechanism that realizes the decoupling. We then establish the actuation modeling and body dynamics to analyze the locomotion behavior. Furthermore, we propose a combined feedforward-feedback control strategy to achieve independent regulation of propulsion and steering. Finally, we validate the feasibility of the design, modeling, and control through a series of prototype experiments, demonstrating swimming, turning, and directional control.

cs.RO↗

arXiv2026

CoFL: Continuous Flow Fields for Language-Conditioned Navigation

Existing language-conditioned navigation systems typically rely on modular pipelines or trajectory generators, but the latter use each scene--instruction annotation mainly to supervise one start-conditioned rollout. To address these limitations, we present CoFL, an end-to-end policy that maps a bird's-eye view (BEV) observation and a language instruction to a continuous flow field for navigation. CoFL reformulates navigation as workspace-conditioned field learning rather than start-conditioned trajectory prediction: it learns local motion vectors at arbitrary BEV locations, turning each scene--instruction annotation into dense spatial control supervision. Trajectories are generated from any start by numerical integration of the predicted field, enabling simple real-time rollout and closed-loop recovery. To enable large-scale training and evaluation, we build a dataset of over 500k BEV image--instruction pairs, each procedurally annotated with a flow field and a trajectory derived from semantic maps built on Matterport3D and ScanNet. Evaluating on strictly unseen scenes, CoFL significantly outperforms modular Vision-Language Model (VLM)-based planners and trajectory generation policies in both navigation precision and safety, while maintaining real-time inference. Finally, we deploy CoFL zero-shot in real-world experiments with BEV observations across multiple layouts, maintaining feasible closed-loop control and a high success rate.

cs.RO↗

arXiv2026

Learning Agile and Robust Omnidirectional Aerial Motion on Overactuated Tiltable-Quadrotors

Tilt-rotor aerial robots enable omnidirectional maneuvering through thrust vectoring, but introduce significant control challenges due to the strong coupling between joint and rotor dynamics. While model-based controllers can achieve high motion accuracy under nominal conditions, their robustness and responsiveness often degrade in the presence of disturbances and modeling uncertainties. This work investigates reinforcement learning for omnidirectional aerial motion control on over-actuated tiltable quadrotors that prioritizes robustness and agility. We present a learning-based control framework that enables efficient acquisition of coordinated rotor-joint behaviors for reaching target poses in the $SE(3)$ space. To achieve reliable sim-to-real transfer while preserving motion accuracy, we integrate system identification with minimal and physically consistent domain randomization. Compared with a state-of-the-art NMPC controller, the proposed method achieves comparable six-degree-of-freedom pose tracking accuracy, while demonstrating superior robustness and generalization across diverse tasks, enabling zero-shot deployment on real hardware.

cs.RO↗

arXiv2026

Hierarchical Trajectory Planning of Floating-Base Multi-Link Robot for Maneuvering in Confined Environments

Floating-base multi-link robots can change their shape during flight, making them well-suited for applications in confined environments such as autonomous inspection and search and rescue. However, trajectory planning for such systems remains an open challenge because the problem lies in a high-dimensional, constraint-rich space where collision avoidance must be addressed together with kinematic limits and dynamic feasibility. This work introduces a hierarchical trajectory planning framework that integrates global guidance with configuration-aware local optimization. First, we exploit the dual nature of these robots - the root link as a rigid body for guidance and the articulated joints for flexibility - to generate global anchor states that decompose the planning problem into tractable segments. Second, we design a local trajectory planner that optimizes each segment in parallel with differentiable objectives and constraints, systematically enforcing kinematic feasibility and maintaining dynamic feasibility by avoiding control singularities. Third, we implement a complete system that directly processes point-cloud data, eliminating the need for handcrafted obstacle models. Extensive simulations and real-world experiments confirm that this framework enables an articulated aerial robot to exploit its morphology for maneuvering that rigid robots cannot achieve. To the best of our knowledge, this is the first planning framework for floating-base multi-link robots that has been demonstrated on a real robot to generate continuous, collision-free, and dynamically feasible trajectories directly from raw point-cloud inputs, without relying on handcrafted obstacle models.

cs.RO↗

arXiv2026

DataChef: Cooking Up Optimal Data Recipes for LLM Adaptation via Reinforcement Learning

In the current landscape of Large Language Models (LLMs), the curation of large-scale, high-quality training data is a primary driver of model performance. A key lever is the \emph{data recipe}, which comprises a data processing pipeline to transform raw sources into training corpora. Despite the growing use of LLMs to automate individual data processing steps, such as data synthesis and filtering, the overall design of data recipes remains largely manual and labor-intensive, requiring substantial human expertise and iteration. To bridge this gap, we formulate \emph{end-to-end data recipe generation} for LLM adaptation. Given a target benchmark and a pool of available data sources, a model is required to output a complete data recipe that adapts a base LLM to the target task. We present DataChef-32B, which performs online reinforcement learning using a proxy reward that predicts downstream performance for candidate recipes. Across six held-out tasks, DataChef-32B produces recipes that yield performance comparable to those curated by human experts. Notably, the recipe from DataChef-32B adapts Qwen3-1.7B-Base to the math domain, achieving 66.7 on AIME'25 and surpassing the official post-training checkpoint (Qwen3-1.7B). This work sheds new light on automating LLM training and developing self-evolving AI systems.

cs.CL↗

arXiv2026

From Failure to Mastery: Generating Hard Samples for Tool-use Agents

The advancement of LLM agents with tool-use capabilities requires diverse and complex training corpora. Existing data generation methods, which predominantly follow a paradigm of random sampling and shallow generation, often yield simple and homogeneous trajectories that fail to capture complex, implicit logical dependencies. To bridge this gap, we introduce HardGen, an automatic agentic pipeline designed to generate hard tool-use training samples with verifiable reasoning. Firstly, HardGen establishes a dynamic API Graph built upon agent failure cases, from which it samples to synthesize hard traces. Secondly, these traces serve as conditional priors to guide the instantiation of modular, abstract advanced tools, which are subsequently leveraged to formulate hard queries. Finally, the advanced tools and hard queries enable the generation of verifiable complex Chain-of-Thought (CoT), with a closed-loop evaluation feedback steering the continuous refinement of the process. Extensive evaluations demonstrate that a 4B parameter model trained with our curated dataset achieves superior performance compared to several leading open-source and closed-source competitors (e.g., GPT-5.2, Gemini-3-Pro and Claude-Opus-4.5). Our code, models, and dataset will be open-sourced to facilitate future research.

cs.CL↗

arXiv2025

FunReason-MT Technical Report: Advanced Data Synthesis Solution for Real-world Multi-Turn Tool-use

Function calling (FC) empowers large language models (LLMs) and autonomous agents to interface with external tools, a critical capability for solving complex, real-world problems. As this ability becomes increasingly central to advanced AI systems, the need for high-quality, multi-turn training data to develop and refine it cannot be overstated. Existing data synthesis methods, such as random environment sampling or multi-agent role-playing, are not powerful enough to generate high-quality data in real-world environments. Practical challenges come in three folds: targeted data synthesis, hard query construction, and multi-turn logical dependency. To address these structural deficiencies, we present FunReason-MT, a novel data synthesis framework for real-world multi-turn tool use. FunReason-MT resolves the complexity barrier in multi-turn FC data by employing 1) Environment-API Graph Interactions to gather varied high-quality trajectories with targeted tool, 2) Advanced Tool-Query Synthesis to simplify hard query construction, and 3) Guided Iterative Chain for sophisticated CoT generation. Evaluations on Berkeley Function-Calling Leaderboard (BFCLv3) demonstrate the power of our framework: a 4B model built upon FunReason-MT generated data achieves state-of-the-art performance among comparable-sized models. Further performance improvements on BFCLv4 confirm that FunReason-MT provides a reliable and robust source for agentic learning.

cs.AI↗

arXiv2025

Benchmarking foundation potentials against quantum chemistry methods for predicting molecular redox potentials

Computational high-throughput virtual screening is essential for identifying redox-active molecules for sustainable applications such as electrochemical carbon capture. A primary challenge in this approach is the high computational cost associated with accurate quantum chemistry calculations. Machine learning foundation potentials (FPs) trained on extensive density functional theory (DFT) calculations offer a computationally efficient alternative. Here, we benchmark the MACE-OMol-0 and UMA FPs against a hierarchy of DFT functionals for predicting experimental molecular redox potentials for both electron transfer (ET) and proton-coupled electron transfer (PCET) reactions. We find that these FPs achieve exceptional accuracy for PCET processes, rivaling their target DFT method. However, the performance is diminished for ET reactions, particularly for multi-electron transfers involving reactive ions that are underrepresented in the OMol25 training data, revealing a key out-of-distribution limitation. To overcome this, we propose an optimal hybrid workflow that uses the FPs for efficient geometry optimization and thermochemical analysis, followed by a crucial single-point DFT energy refinement and an implicit solvation correction. This pragmatic approach provides a robust and scalable strategy for accelerating high-throughput virtual screening in sustainable chemistry.

physics.chem-ph↗

arXiv2025

SimPoly: Simulation of Polymers with Machine Learning Force Fields Derived from First Principles

Polymers are a versatile class of materials with widespread industrial applications. Advanced computational tools could revolutionize their design, but their complex, multi-scale nature poses significant modeling challenges. Conventional force fields often lack the accuracy and transferability required to capture the intricate interactions governing polymer behavior. Conversely, quantum-chemical methods are computationally prohibitive for the large systems and long timescales required to simulate relevant polymer phenomena. Here, we overcome these limitations with a machine learning force field (MLFF) approach. We demonstrate that macroscopic properties for a broad range of polymers can be predicted ab initio, without fitting to experimental data. Specifically, we develop a fast and scalable MLFF to accurately predict polymer densities, outperforming established classical force fields. Our MLFF also captures second-order phase transitions, enabling the prediction of glass transition temperatures. To accelerate progress in this domain, we introduce a benchmark of experimental bulk properties for 130 polymers and an accompanying quantum-chemical dataset. This work lays the foundation for a fully in silico design pipeline for next-generation polymeric materials.

physics.chem-ph↗

arXiv2025

Reasoning through Exploration: A Reinforcement Learning Framework for Robust Function Calling

The effective training of Large Language Models (LLMs) for function calling faces a critical challenge: balancing exploration of complex reasoning paths with stable policy optimization. Standard methods like Supervised Fine-Tuning (SFT) fail to instill robust reasoning, and traditional Reinforcement Learning (RL) struggles with inefficient exploration. We propose \textbf{EGPO}, a new RL framework built upon Group Relative Policy Optimization (GRPO), designed to address this challenge directly. The core of EGPO is an entropy-enhanced advantage function that integrates the entropy of the model's Chain-of-Thought (CoT) into the policy gradient computation. This encourages the generation of diverse reasoning strategies. To maintain optimization direction, the entropy bonus is carefully constrained by a clipping mechanism. Complemented by a strict, binary reward signal, EGPO effectively guides the model towards discovering structured and accurate tool invocation patterns. On the challenging Berkeley Function Calling Leaderboard (BFCL), a 4B-parameter model trained with EGPO sets a new state-of-the-art among models of comparable size, surpassing a range of strong competitors, including GPT-4o and Gemini-2.5.

cs.LG↗

arXiv2025

Hierarchical Language Models for Semantic Navigation and Manipulation in an Aerial-Ground Robotic System

Heterogeneous multirobot systems show great potential in complex tasks requiring coordinated hybrid cooperation. However, existing methods that rely on static or task-specific models often lack generalizability across diverse tasks and dynamic environments. This highlights the need for generalizable intelligence that can bridge high-level reasoning with low-level execution across heterogeneous agents. To address this, we propose a hierarchical multimodal framework that integrates a prompted large language model (LLM) with a fine-tuned vision-language model (VLM). At the system level, the LLM performs hierarchical task decomposition and constructs a global semantic map, while the VLM provides semantic perception and object localization, where the proposed GridMask significantly enhances the VLM's spatial accuracy for reliable fine-grained manipulation. The aerial robot leverages this global map to generate semantic paths and guide the ground robot's local navigation and manipulation, ensuring robust coordination even in target-absent or ambiguous scenarios. We validate the framework through extensive simulation and real-world experiments on long-horizon object arrangement tasks, demonstrating zero-shot adaptability, robust semantic navigation, and reliable manipulation in dynamic environments. To the best of our knowledge, this work presents the first heterogeneous aerial-ground robotic system that integrates VLM-based perception with LLM-driven reasoning for global high-level task planning and execution.

cs.RO↗