Searcharxiv⌕ Search

arXiv subjects

Zixi Chen

Publications and source records attributed to Zixi Chen.

At least 19 recordsLinked to original sources
arXiv2026

StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability?

Humans need to study only a handful of well-written textbooks to master a discipline and attempt its hardest problems. We argue that an ideal self-evolution method should share the same property, that is autonomously learning from raw training material for transferable problem-solving capability. However, we still lack a direct measurement for it. We introduce StudyBench, a controlled physics benchmark that directly measures how efficiently a self-evolution method converts training material into capability. We organise the test set into an Application Set, consisting of difficult textbook problems and evaluating absorption ability, and a Transfer Set, consisting of olympiad-level problems and evaluating transfer ability. Benchmarking representative self-evolution methods across three base models, we find that improvements on the Application Set rarely translate to the harder Transfer Set. A guidance ablation exposes a Guidance Gap: even the strongest method closes only a small fraction of what the same material unlocks when supplied as in-context guidance. Besides, every method hits a Compute Plateau, saturating well before exhausting its compute budget. The remaining gap is therefore a method problem rather than a data or compute problem. By offering a clean and controlled benchmark, StudyBench turns self-evolution progress from an open-ended pursuit into a measurable target for future research. Our code is released at https://github.com/thunlp/StudyBench.

cs.AI↗
arXiv2026

Refined Thompson Learning for Adaptive Bandits: Power-Efficient Flexibility Scheduling Across Data Centers

The rapid growth of large-scale AI workloads in data centers has placed increasing pressure on power grids in recent years. Since power systems must continuously balance supply and demand, there is growing interests in leveraging data-center workload flexibility as a grid service. We propose a contextual restless multi-armed bandit (CRMAB) framework in which a grid operator requests load reductions without observing internal job-scheduling decisions. Under index-ability guarantee, each data center or physical machine is modeled as a Markov decision process (MDP) over a cyclic virtual-machine (VM) job queue, with unknown rewards and transition dynamics learned online using Thompson sampling and Whittle-index policies. To improve learning under sparse and noisy observations, the framework augments an adaptive Thompson--Whittle (TW) policy with domain-informed transition priors and gated prior mixing. In baseline experiments, the best adaptive refined variant achieves 91.4\% of the oracle reward after 100 rounds and 96.8\% after 1,000 rounds. Across a 16-setting stress test spanning different state-space sizes and levels of contextual noise, the best refined variant consistently outperforms the original TW policy with high confidence while remaining competitive with EXP4. A graph-based prior further incorporates data-center hardware constraints, including computing-resource limits. Overall, the results demonstrate the economic potential of data-center flexibility as a grid service and highlight the importance of high-quality, open-source AI workload traces for developing and evaluating such services.

cs.CE↗
arXiv2026

Online Linear Programming for Multi-Objective Routing in LLM Serving

We study the online routing problem in large language model serving, where requests arrive sequentially and must be dispatched to parallel decode workers under tight batch-size and KV-cache constraints. Unlike widely used routing heuristics that are not tied to explicit service-level objectives (SLOs) and offer limited control over latency-throughput trade-offs, we introduce a multi-objective optimization framework that formulates routing as an online linear programming with interpretable decision rewards. We apply an efficient bid-price control policy based on the online linear programming that admits requests when their SLO-weighted benefit exceeds their shadow prices. To meet millisecond decision requirements, we develop a warm-started, projected first-order updates that track the evolving dual shadow prices online with predictable runtime. We integrate our router into the Vidur simulator and demonstrate substantial improvements over standard baselines across multiple SLO regimes, including end-to-end latency, time-to-first-token, throughput, and tail performance. A big picture from our result: a science-based approach outperforms others based on heuristics.

cs.AI↗
arXiv2026

Strouhal-Aware Model Predictive Control for Efficient Multi-Fin Flapping Locomotion

Efficient flapping propulsion hinges on operating within a narrow Strouhal number window, a principle nature has converged upon for maximum thrust-to-power ratio. We translate this bioinspired empirical rule into real-time control, demonstrating it on an autonomous underwater vehicle driven by four soft fins. The proposed Strouhal-aware Model Predictive Control (MPC) enhances a quasi-steady hydrodynamic model with an explicit penalty for Strouhal deviation, solving the resulting nonconvex problem via a two-stage sampling and gradient optimization that runs onboard at 25 Hz. Pool and field trials show that the controller keeps each fin within the optimal Strouhal corridor (0.25-0.35) while precisely tracking commanded forces. This results in a mean reduction in mechanical power of 8.8\% to 32\% throughout the cruising range of 0.1 to 0.3 m/s. The proposed method also allows for a velocity of 0.4 m/s, which is unattainable for a baseline of the conventional inverse model. The results confirm that embedding first-principle flow physics into an MPC objective yields tangible endurance gains without sacrificing agility, offering a generic pathway to energy-aware locomotion in next-generation multifin robots.

cs.RO↗
arXiv2026

Emergent Capabilities Arise Randomly from Learning Sparse Attention Patterns

Neural scaling laws for transformer language models predict smooth improvements in pretraining loss with increasing parameters, but downstream capabilities such as in-context learning are known to emerge abruptly past a certain model scale. In this paper, we show that emergent capabilities arise stochastically throughout training, with larger models acquiring them earlier on average. We demonstrate that the emergence of capabilities such as pattern completion and indirect object identification corresponds to the abrupt learning of task-relevant attention patterns. To isolate this phenomenon, we train transformer models on synthetic linear map and cellular automata datasets, and we show that the difficulty of learning attention patterns depends on context length and pattern sparsity. Moreover, scaling the number of attention heads improves learning efficiency on our synthetic tasks, while increasing the head dimension yields diminishing returns past a minimum capacity. We additionally investigate architectures with alternative attention mechanisms, showing that MLP-Mixer outperforms a transformer on linear map tasks with complex attention patterns. Our findings provide a mechanistic insight into emergence, showing that downstream capabilities arise abruptly due to the intrinsic difficulty of learning sparse attention patterns in transformer models.

cs.LG↗
arXiv2026

Robust Restless Multi-Armed Bandit for Data Center Flexibility Services Through Virtual Machine Scheduling

Energy demands from data centers have surged and stressed the grid in recent years. Electric grids require balancing supply and demand every second, motivating demand response (reduction) from large loads, including data centers. This can be achieved by rescheduling jobs on a physical machine. Its real-time implementation is uncertain due to fluctuating resource utilization, and rescheduling incurs quality-of-service (QoS) losses that providers are unwilling to disclose. We propose a restless multi-armed bandit (RMAB) framework, in which the grid operator requests load reductions without access to detailed job-rescheduling procedures. Using open-source virtual machine (VM) datasets, we model job arrivals and rescheduling at each data center as a restless arm in a Markov decision process (MDP) and derive Whittle-index-based policies using the learned transition function via Thompson sampling. To overcome the weakness of an increasingly long learning process due to an enlarged state space, we use a mixed strategy that includes a global upper confidence bound (UCB) and encodes trust indices to enhance robustness and accelerate learning. Results show that the proposed mixed-strategy algorithm remains robust across varying state-space sizes and consistently outperforms the pure Thompson-Whittle (TW) algorithm, especially when contextual information is noisy. It also demonstrates superior performance compared to the state-of-the-art EXP4 framework. We provided open-source code to ensure reproducibility.

cs.CE↗
arXiv2026

Tackling the Data-Parallel Load Balancing Bottleneck in LLM Serving: Practical Online Routing at Scale

Data-parallel (DP) load balancing has emerged as a first-order bottleneck in large-scale LLM serving. When a model is sharded across devices via tensor parallelism (TP) or expert parallelism (EP) and replicated across many DP workers, every decode step ends in a synchronization barrier whose latency is set by the most heavily loaded worker; even modest persistent imbalance across DP workers compounds, step after step, into a substantial fraction of wasted compute. The problem is hard for reasons specific to LLM decoding: assignments are sticky (migrating KV caches has a high cost), per-request loads grow over time, arrivals are non-stationary, and the router must decide within a sub-100\,ms decode budget over hundreds of waiting requests and tens of workers. We present \textbf{BalanceRoute}, a family of practical online routing algorithms that target this bottleneck. The first, \textbf{BR-0}, requires no prediction infrastructure and uses a piecewise-linear F-score that captures the sharp asymmetry between admissions that fill safe margin and those that overflow into the envelope; a two-stage decomposition keeps per-step cost compatible with millisecond-scale scheduling. The second, \textbf{BR-H}, generalizes BR-0 with a short, constant lookahead $H$ and a lightweight termination-classifier interface, extending the F-score to a horizon-discounted form. We deploy BalanceRoute on a 144-NPU cluster and evaluate against vLLM baselines on both a proprietary production trace and the public Azure-2024 trace. Across both workloads, BalanceRoute substantially reduces average DP imbalance and improves end-to-end serving throughput.

cs.DC↗
arXiv2026

A Flow Matching Framework for Soft-Robot Inverse Dynamics

Learning the inverse dynamics of soft continuum robots remains challenging due to high-dimensional nonlinearities and complex actuation coupling. Conventional feedback-based controllers often suffer from control chattering due to corrective oscillations, while deterministic regression-based learners struggle to capture the complex nonlinear mappings required for accurate dynamic tracking. Motivated by these limitations, we propose an inverse-dynamics framework for open-loop feedforward control that learns the system's differential dynamics as a generative transport map. Specifically, inverse dynamics is reformulated as a conditional flow-matching problem, and Rectified Flow (RF) is adopted as a lightweight instance to generate physically consistent control inputs rather than conditional averages. Two variants are introduced to further enhance physical consistency: RF-Physical, utilizing a physics-based prior for residual modeling; and RF-FWD, integrating a forward-dynamics consistency loss during flow matching. Extensive evaluations demonstrate that our framework reduces trajectory tracking RMSE by over 50% compared to standard regression baselines (MLP, LSTM, Transformer). The system sustains stable open-loop execution at a peak end-effector velocity of 1.14 m/s with sub-millisecond inference latency (0.995 ms). This work demonstrates flow matching as a robust, high-performance paradigm for learning differential inverse dynamics in soft robotic systems.

cs.RO↗
arXiv2026

PlugMem: A Task-Agnostic Plugin Memory Module for LLM Agents

Long-term memory is essential for large language model (LLM) agents operating in complex environments, yet existing memory designs are either task-specific and non-transferable, or task-agnostic but less effective due to low task-relevance and context explosion from raw memory retrieval. We propose PlugMem, a task-agnostic plugin memory module that can be attached to arbitrary LLM agents without task-specific redesign. Motivated by the fact that decision-relevant information is concentrated as abstract knowledge rather than raw experience, we draw on cognitive science to structure episodic memories into a compact, extensible knowledge-centric memory graph that explicitly represents propositional and prescriptive knowledge. This representation enables efficient memory retrieval and reasoning over task-relevant knowledge, rather than verbose raw trajectories, and departs from other graph-based methods like GraphRAG by treating knowledge as the unit of memory access and organization instead of entities or text chunks. We evaluate PlugMem unchanged across three heterogeneous benchmarks (long-horizon conversational question answering, multi-hop knowledge retrieval, and web agent tasks). The results show that PlugMem consistently outperforms task-agnostic baselines and exceeds task-specific memory designs, while also achieving the highest information density under a unified information-theoretic analysis. Code and data are available at https://github.com/TIMAN-group/PlugMem.

cs.CL↗
arXiv2026

Deep-Learning-Based Control of a Decoupled Two-Segment Continuum Robot for Endoscopic Submucosal Dissection

Manual endoscopic submucosal dissection (ESD) is technically demanding, and existing single-segment robotic tools offer limited dexterity. These limitations motivate the development of more advanced solutions. To address this, DESectBot, a novel dual segment continuum robot with a decoupled structure and integrated surgical forceps, enabling 6 degrees of freedom (DoFs) tip dexterity for improved lesion targeting in ESD, was developed in this work. Deep learning controllers based on gated recurrent units (GRUs) for simultaneous tip position and orientation control, effectively handling the nonlinear coupling between continuum segments, were proposed. The GRU controller was benchmarked against Jacobian based inverse kinematics, model predictive control (MPC), a feedforward neural network (FNN), and a long short-term memory (LSTM) network. In nested-rectangle and Lissajous trajectory tracking tasks, the GRU achieved the lowest position/orientation RMSEs: 1.11 mm/ 4.62{\deg} and 0.81 mm/ 2.59{\deg}, respectively. For orientation control at a fixed position (four target poses), the GRU attained a mean RMSE of 0.14 mm and 0.72{\deg}, outperforming all alternatives. In a peg transfer task, the GRU achieved a 100% success rate (120 success/120 attempts) with an average transfer time of 11.8s, the STD significantly outperforms novice-controlled systems. Additionally, an ex vivo ESD demonstration grasping, elevating, and resecting tissue as the scalpel completed the cut confirmed that DESectBot provides sufficient stiffness to divide thick gastric mucosa and an operative workspace adequate for large lesions.These results confirm that GRU-based control significantly enhances precision, reliability, and usability in ESD surgical training scenarios.

cs.RO↗
arXiv2026

ProjDevBench: Benchmarking AI Coding Agents on End-to-End Project Development

Recent coding agents can generate complete codebases from simple prompts, yet existing evaluations focus on issue-level bug fixing and lag behind end-to-end development. We introduce ProjDevBench, an end-to-end benchmark that provides project requirements to coding agents and evaluates the resulting repositories. Combining Online Judge (OJ) testing with LLM-assisted code review, the benchmark evaluates agents on (1) system architecture design, (2) functional correctness, and (3) iterative solution refinement. We curate 20 programming problems across 8 categories, covering both concept-oriented tasks and real-world application scenarios, and evaluate six coding agents built on different LLM backends. Our evaluation reports an overall acceptance rate of 27.38%: agents handle basic functionality and data structures but struggle with complex system design, time complexity optimization, and resource management. Our benchmark is available at https://github.com/zsworld6/projdevbench.

cs.AI↗
arXiv2026

Analytical Provisioning for Attention-FFN Disaggregated LLM Serving under Stochastic Workloads

Attentio-FFN disaggregation (AFD) is an emerging architecture for LLM decoding that separates state-heavy, KV-cache-dominated Attention computation from stateless, compute-intensive FFN computation, connected by per-step communication. While AFD enables independent scaling of memory and compute resources, its performance is highly sensitive to the Attention/FFN provisioning ratio: mis-sizing induces step-level blocking and costly device idle time. We develop an analytical provisioning framework for AFD bundles in an $r$A--$1$F topology under stochastic workloads. Two sources of randomness shape the problem: per-slot Attention workload evolves as KV caches grow and completed requests are replenished with random prompt and decode lengths, and synchronized execution across Attention workers introduces a barrier governed by the slowest worker. We address both via a renewal-reward characterization of the per-slot stationary token load, identifying a single workload statistic $\theta$ that governs provisioning under arbitrary prefill-decode distributions and admits a nonparametric estimator from request traces. The analysis yields a closed-form mean-field rule for the optimal A/F ratio decomposing into Attention-, communication-, and FFN-bottleneck regimes, together with a Gaussian barrier-aware refinement that quantifies cross-worker synchronization overhead. A trace-calibrated AFD simulator supports the framework across workloads: the predicted optimal ratio matches the simulation-optimal within 10%. Together, these results provide a compact, calibratable account of how stochastic workload structure determines provisioning in disaggregated LLM serving.

cs.LG↗
arXiv2026

A Universal Load Balancing Principle and Its Application to Large Language Model Serving

Over 40% of computational power in Large Language Model (LLM) serving systems can be systematically wasted - not from hardware limits, but from load imbalance in barrier-synchronized parallel processing. When progress is gated by the slowest worker at each step, heterogeneous and evolving workloads create persistent stragglers; faster workers idle while drawing power, producing nothing. In large language model inference alone, this translates to gigawatt-hours of wasted electricity daily. Here we develop a universal load-balancing principle for barrier-synchronized systems with non-migratable state. We prove worst-case theoretical guarantees: imbalance reduction grows with system scale, and the resulting energy savings can exceed 52% for modern hardware at fleet scale. Experiments corroborate the theory, demonstrating 28% energy reduction alongside substantial throughput and latency improvements. Formulated as an online integer optimization with provable guarantees, the principle extends beyond LLM serving to broad classes of barrier-synchronized parallel systems, establishing a theoretical foundation for sustainable high-performance computing.

cs.DC↗
arXiv2026

OrthoGeoLoRA: Geometric Parameter-Efficient Fine-Tuning for Structured Social Science Concept Retrieval on theWeb

Large language models and text encoders increasingly power web-based information systems in the social sciences, including digital libraries, data catalogues, and search interfaces used by researchers, policymakers, and civil society. Full fine-tuning is often computationally and energy intensive, which can be prohibitive for smaller institutions and non-profit organizations in the Web4Good ecosystem. Parameter-Efficient Fine-Tuning (PEFT), especially Low-Rank Adaptation (LoRA), reduces this cost by updating only a small number of parameters. We show that the standard LoRA update $\Delta W = BA^\top$ has geometric drawbacks: gauge freedom, scale ambiguity, and a tendency toward rank collapse. We introduce OrthoGeoLoRA, which enforces an SVD-like form $\Delta W = B\Sigma A^\top$ by constraining the low-rank factors to be orthogonal (Stiefel manifold). A geometric reparameterization implements this constraint while remaining compatible with standard optimizers such as Adam and existing fine-tuning pipelines. We also propose a benchmark for hierarchical concept retrieval over the European Language Social Science Thesaurus (ELSST), widely used to organize social science resources in digital repositories. Experiments with a multilingual sentence encoder show that OrthoGeoLoRA outperforms standard LoRA and several strong PEFT variants on ranking metrics under the same low-rank budget, offering a more compute- and parameter-efficient path to adapt foundation models in resource-constrained settings.

cs.CL↗
arXiv2025

Hyperparameter Transfer Enables Consistent Gains of Matrix-Preconditioned Optimizers Across Scales

Several recently introduced deep learning optimizers utilizing matrix-level preconditioning have shown promising speedups relative to the current dominant optimizer AdamW, particularly in relatively small-scale experiments. However, efforts to validate and replicate their successes have reported mixed results. To better understand the effectiveness of these optimizers at scale, in this work we investigate how to scale preconditioned optimizers via hyperparameter transfer, building on prior works such as $\mu$P. We study how the optimal learning rate and weight decay should scale with model width and depth for a wide range of optimizers, including Shampoo, SOAP, and Muon, accounting for the impact of commonly used techniques such as blocking and grafting. We find that scaling the learning rate according to $\mu$P improves transfer, but can still suffer from significant finite-width deviations that cause drifting optimal learning rates, which we show can be mitigated by blocking and explicit spectral normalization. For compute-optimal scaling, we find scaling independent weight decay as $1/\mathrm{width}$ is nearly optimal across optimizers. Applying these scaling rules, we show Muon, SOAP and Shampoo consistently achieve near $1.4\times$ speedup over AdamW for training Llama-architecture language models of sizes ranging from $190$M to $1.4$B, whereas the speedup vanishes rapidly with scale under incorrect scaling. Based on these results and further ablations, we argue that studying optimal hyperparameter transfer is essential for reliably comparing optimizers at scale given a realistic tuning budget.

cs.LG↗
arXiv2025

AdapJ: An Adaptive Extended Jacobian Controller for Soft Manipulators

The nonlinearity and hysteresis of soft robot motions present challenges for control. To solve these issues, the Jacobian controller has been applied to approximate the nonlinear behaviors in a linear format. Accurate controllers like neural networks can handle delayed and nonlinear motions, but they require large datasets and exhibit low adaptability. Based on a novel analysis on these controllers, we propose an adaptive extended Jacobian controller, AdapJ, for soft manipulators. This controller retains the concise format of the Jacobian controller but introduces independent parameters. Similar to neural networks, its initialization and updating mechanism leverages the inverse model without building the corresponding forward model. In the experiments, we first compare the performance of the Jacobian controller, model predictive controller, neural network controller, iterative feedback controller, and AdapJ in simulation. We further analyze how AdapJ parameters adapt in response to the physical property change. Then, real-world experiments have validated that AdapJ outperforms the neural network controller, model predictive controller, and iterative feedback controller with fewer training samples and adapts robustly to varying conditions, including different control frequencies, material softness, and external disturbances. Future work may include online adjustment of the controller format and adaptability validation in more scenarios.

cs.RO↗
arXiv2025

Convergence Rate in Nonlinear Two-Time-Scale Stochastic Approximation with State (Time)-Dependence

The nonlinear two-time-scale stochastic approximation is widely studied under conditions of bounded variances in noise. Motivated by recent advances that allow for variability linked to the current state or time, we consider state- and time-dependent noises. We show that the Lyapunov function exhibits polynomial convergence rates in both cases, with the rate of polynomial delay depending on the parameters of state- or time-dependent noises. Notably, if the state noise parameters fully approach their limiting value, the Lyapunov function achieves an exponential convergence rate. We provide two numerical examples to illustrate our theoretical findings in the context of stochastic gradient descent with Polyak-Ruppert averaging and stochastic bilevel optimization.

math.OC↗
arXiv2025

Adaptively Robust LLM Inference Optimization under Prediction Uncertainty

We study the problem of optimizing Large Language Model (LLM) inference scheduling to minimize total latency. LLM inference is an online and multi-task service process and also heavily energy consuming by which a pre-trained LLM processes input requests and generates output tokens sequentially. Therefore, it is vital to improve its scheduling efficiency and reduce the power consumption while a great amount of prompt requests are arriving. A key challenge in LLM inference scheduling is that while the prompt length is known upon arrival, the output length, which critically impacts memory usage and processing time, is unknown. To address this uncertainty, we propose algorithms that leverage machine learning to predict output lengths, assuming the prediction provides an interval classification (min-max range) for each request. We first design a conservative algorithm, $\mathcal{A}_{\max}$, which schedules requests based on the upper bound of predicted output lengths to prevent memory overflow. However, this approach is overly conservative: as prediction accuracy decreases, performance degrades significantly due to potential overestimation. To overcome this limitation, we propose $\mathcal{A}_{\min}$, an adaptive algorithm that initially treats the predicted lower bound as the output length and dynamically refines this estimate during inferencing. We prove that $\mathcal{A}_{\min}$ achieves a log-scale competitive ratio. Through numerical simulations, we demonstrate that $\mathcal{A}_{\min}$ often performs nearly as well as the hindsight scheduler, highlighting both its efficiency and robustness in practical scenarios. Moreover, $\mathcal{A}_{\min}$ relies solely on the lower bound of the prediction interval--an advantageous design choice since upper bounds on output length are typically more challenging to predict accurately.

cs.LG↗