arXiv subjects

Yize Chen

Publications and source records attributed to Yize Chen.

At least 19 recordsLinked to original sources

arXiv2026

Beyond Success and Failure: Length-Aware Contrastive Learning for GUI Agents

Graphical User Interface (GUI) agents powered by Multimodal Large Language Models (MLLMs) have shown strong potential for automating tasks across diverse digital environments, where reinforcement learning (RL) has become a dominant training paradigm. However, widely used methods such as Group Relative Policy Optimization (GRPO) suffer from reward-gradient misalignment, leading to inefficient and unstable optimization. Recent work addresses this issue by reformulating RL with verifiable rewards (RLVR) as contrastive or classification-based objectives, which improve stability by eliminating problematic gradient behaviors. Despite this progress, existing contrastive RLVR methods rely primarily on outcome-level supervision and fail to capture fine-grained differences in trajectory quality within the same outcome category. In this paper, we propose Length-Aware Contrastive Learning for GUI Agents (LACL-GUI), a contrastive RLVR framework that incorporates trajectory-level quality signals into policy optimization. LACL-GUI introduces structured preferences within both successful and failed trajectories, encouraging concise successful executions and differentiating failure quality based on divergence from successful trajectories, while preserving optimization stability. Experiments on GUI agent benchmarks show that LACL-GUI provides more effective learning signals and consistently improves agent performance over prior methods, highlighting the value of trajectory-level supervision in contrastive RLVR.

cs.AI↗

arXiv2026

Beyond Imitation: Filtering On-Policy Distillation by Reasoning Progress

On-policy distillation (OPD) has emerged as an effective framework for post-training language models by pairing student-generated trajectories with dense token-level supervision from a teacher. However, OPD implicitly assumes that teacher-derived rewards are an appropriate proxy for reasoning progress, and therefore treats all teacher feedback equally during policy optimization. While in practice, this assumption does not always hold. We observe that teacher-derived rewards often conflict with genuine reasoning progress, as reasoning steps with clear reasoning advancement may still receive lower distillation rewards, simply due to deviation from teacher's outputs. To address this mismatch, we propose Reasoning-Progress-Aware Reward Filtering for On-Policy Distillation (R2-OPD), which constructs two within-trajectory rankings of reasoning spans, one from teacher-derived rewards and the other from independently estimated progress reward. Distillation rewards are selectively suppressed whenever the two rankings disagree, reducing supervision that conflicts with reasoning progress while preserving effective teacher guidance. Our approach shows consistent improvement over standard OPD especially regarding reasoning performances.

cs.AI↗

arXiv2026

Smoothing the Ramp, Not the Peak: Scheduling-Induced Power Dynamics of LLM Inference and Their Grid-Scale Consequences

Large language model (LLM) inference serving is a fast-growing electricity load whose power dynamics remain uncharacterized from a grid-planning perspective. Using real, measured GPU power traces, we show that chunked prefill scheduling, a latency-motivated technique already deployed by default in production LLM serving, is a controllable knob that regulates power ramp rate without touching peak power. Contrary to the intuitive hypothesis that splitting a long prompt's computation into smaller steps should flatten its power spike, peak power stays relatively the same while mean ramp rate falls substantially. Critically, this ramp-rate benefit is not a fixed property of the policy: it grows monotonically with system saturation, and we confirm this along two independent axes: concurrency (7.0% at light load to 34.6% at heavy load, mean-ramp reduction) and long-prompt ("whale") request load (from statistically flat at low whale incidence to 42.6% at high whale fraction/size). We translate this single-GPU mechanism into an operational grid quantity, regulation-reserve procurement, posed and solved as a chance-constrained problem using a model-free bootstrap directly resampling real measured power traces. At a representative operating point, this translates to an estimated 20.3-22.7% reduction in the fast-ramping reserve capacity a grid operator would need to provision, across reliability levels from 95% to 99.9%. Together, these results give grid operators a no-cost demand-shaping tool available today, whose benefit is largest precisely when data centers run hottest and grid stress is most salient.

eess.SY↗

arXiv2026

LLMs and Agentic AI Systems for Smart Grids: A Tutorial on Architectures and Applications

Large language models (LLMs) and agentic AI systems have evolved from natural language tasks to using external tools to plan, retrieve, and act in technical domains. In smart grids, recent work applies agentic schemes to forecasting, optimization, and control, wrapping trusted solvers behind language interfaces and orchestrating multi-step workflows. The literature lacks a unified approach to designing and evaluating such systems. LLMs can produce numerically plausible yet physically infeasible outputs, evaluation protocols vary across tasks, and the boundary between what the model should and should not compute is implicit. This paper presents a solver-grounded design principle: a numerical result is reported only when it originates from a trusted tool and passes explicit verification. We review the building blocks of LLM and agentic AI systems for power systems: prompting strategies and agentic architectures. We instantiate the principle in four case studies: wind power forecasting, EV charging scheduling, power flow analysis, and contingency diagnosis, each comparing an LLM-only baseline against its solver-grounded counterpart on identical data and metrics. EVAgent reproduces the CVXPY optimum while reducing LLM-only unmet energy by 7.5-9.5x, and GridDebugAgent repairs 17/39 contingency cases while reducing total violations by 52.3%. We propose a four-group evaluation framework spanning task utility, solver-grounded correctness, faithfulness and safe failure, and cost and latency. A consistent division of labor emerges: the agentic system reliably orchestrates, retrieves, and explains, while trusted tools compute and a verification gate decides what is reported.

eess.SY↗

arXiv2026

TimeRFT: Stimulating Generalizable Time Series Forecasting for TSFMs via Reinforcement Finetuning

Time Series Foundation Models (TSFMs) have demonstrated strong generalization capability and data efficiency in time series forecasting through large-scale pretraining. However, adapting TSFMs to downstream forecasting tasks remains challenging due to temporal distribution shifts and varying data availability. Specifically, the non-stationary and uncertain nature of time series data leads to discrepancies between historical training and future forecasting distributions, making existing Supervised FineTuning (SFT)-based adaptation vulnerable to overfitting and limited generalization. Moreover, forecasting tasks often operate under varying data regimes, requiring TSFMs to extract generalizable temporal patterns from limited training samples. To address these challenges, we propose Time series Reinforcement FineTuning (TimeRFT), a reinforcement learning-based adaptation paradigm for TSFMs. TimeRFT introduces two forecasting-oriented training recipes: (i) A quality-aware temporal reward mechanism providing fine-grained credit assignment by holistically evaluating the contribution of each prediction step to overall forecasting performance. (ii) A difficulty-aware data selection strategy prioritizing informative time series samples with generalizable forecasting patterns. Extensive experiments on diverse real-world forecasting benchmarks demonstrate that TimeRFT consistently surpasses SFT-based adaptation methods across various real-world forecasting tasks with different data regimes, achieving improved prediction accuracy and enhanced generalization against unforeseen distribution shifts.

eess.SP↗

arXiv2026

To Defer or To Shift? The Role of AI Data Center Flexibility on Grid Interconnection

The integration of AI data centers into power grid represents one of the most emerging and complex challenges for the energy systems. As computational demand scales at an unprecedented rate, the traditional grid planning study's paradigm of treating data centers as rigid, inflexible loads is becoming economically, mathematically and operationally untenable. This work tries to understand and address the large load interconnection bottleneck by modeling and evaluating AI load flexibility. By examining data center's temporal and spatial shifting capabilities within a grid capacity expansion framework, we build a quantitative grid planning model, and evaluate their impacts on additional generation, operational costs, and network congestion. Numerical study reveals interesting observations, as AI data center flexibility are not felt consistently, and increasing flexibility does not necessarily translate to less generation capacity required. Depending on data center's locations, flexibility range, and grid load conditions, flexible AI load can help reduce grid investment and operational costs by 3-21%. Our work also indicate that longer deferral time of AI compute has diminishing returns for offloading grid electricity dispatch pressure.

eess.SY↗

arXiv2026

A Pontryagin Method of Model-based Reinforcement Learning via Hamiltonian Actor-Critic

Model-based reinforcement learning (MBRL) improves sample efficiency by leveraging learned dynamics models for policy optimization. However, the effectiveness of methods such as actor-critic is often limited by compounding model errors, which degrade long-horizon value estimation. Existing approaches, such as Model-Based Value Expansion (MVE), partially mitigate this issue through multi-step rollouts, but remain sensitive to rollout horizon selection and residual model bias. Motivated by the Pontryagin Maximum Principle (PMP), we propose Hamiltonian Actor-Critic (HAC), a model-based approach that eliminates explicit value function learning by directly optimizing a Hamiltonian defined over the learned dynamics and reward for deterministic systems. By avoiding value approximation, HAC reduces sensitivity to model errors while admitting convergence guarantees. Extensive experiments on continuous control benchmarks, in both online and offline RL settings, demonstrate that HAC outperforms model-free and MVE-based baselines in control performance, convergence speed, and robustness to distributional shift, including out-of-distribution (OOD) scenarios. In offline settings with limited data, HAC matches or exceeds state-of-the-art methods, highlighting its strong sample efficiency.

eess.SY↗

arXiv2026

Switching-Reference Voltage Control for Distribution Systems with AI-Training Data Centers

Large-scale AI training workloads in data centers exhibit rapid and periodic power swings that can induce voltage deviations in power distribution systems. Existing voltage controllers treat these swings as a generic disturbance, leading to high control effort but still large voltage violations. However, such emerging loads are not random: they alternate between two distinct operating phases. This paper exploits this structure with a decentralized switching-reference voltage control framework. By switching each voltage reference in step with the workload phases, the controller cancels the phase-induced voltage shift, holding the voltage within limits with low control effort. Because real-time communication between buses is not always available, the controller is designed to infer the reference from local voltage measurements. This paper further proves convergence under deadband and saturation. In case studies on real AI training power traces, the switching reference suppresses voltage violations, sometimes eliminating them entirely, while reducing the control effort by approximately an order of magnitude compared with conventional droop control. Further experiments confirm that it remains effective with multiple data centers and internal load smoothing.

eess.SY↗

arXiv2026

Online Electric Vehicle Charging Control with Battery Thermal Management in Cold Environments

Electric vehicle (EV) adoption in cold regions is hindered by degraded EV charging performance at low temperatures, which necessitates effective battery thermal management during charging. Given the coupling of battery charging and heating dynamics, this paper examines the benefits of online coordinated charging and heating control, rather than performing them separately. Specifically, we first build queue models for both battery charging and thermal dynamics. Then, we formulate an optimization problem to minimize the system cost of a charging station, which allows us to coordinate charging and heating through maintaining queue stability. To solve the problem, we develop our online coordinated charging and heating control algorithm within the theoretical framework of Lyapunov optimization. Note that our online method is prediction-free and independent of any assumed modeling of uncertainty. We also characterize both the feasibility and optimality of the proposed control approach. Numerical results based on real-world data demonstrate the effectiveness and robustness of our control method through comparisons.

math.OC↗

arXiv2026

STO-RL: Offline RL under Sparse Rewards via LLM-Guided Subgoal Temporal Order

Offline reinforcement learning (RL) enables policy learning from pre-collected datasets, avoiding costly and risky online interactions, but it often struggles with long-horizon tasks involving sparse rewards. Existing goal-conditioned and hierarchical offline RL methods decompose such tasks and generate intermediate rewards to mitigate limitations of traditional offline RL, but usually overlook temporal dependencies among subgoals and rely on imprecise reward shaping, leading to suboptimal policies. To address these issues, we propose STO-RL (Offline RL using LLM-Guided Subgoal Temporal Order), an offline RL framework that leverages large language models (LLMs) to generate temporally ordered subgoal sequences and corresponding state-to-subgoal-stage mappings. Using this temporal structure, STO-RL applies potential-based reward shaping to transform sparse terminal rewards into dense, temporally consistent signals, promoting subgoal progress while avoiding suboptimal solutions. The resulting augmented dataset with shaped rewards enables efficient offline training of high-performing policies. Evaluations on four discrete and continuous sparse-reward benchmarks demonstrate that STO-RL consistently outperforms state-of-the-art offline goal-conditioned and hierarchical RL baselines, achieving faster convergence, higher success rates, and shorter trajectories. Ablation studies further confirm STO-RL's robustness to imperfect or noisy LLM-generated subgoal sequences, demonstrating that LLM-guided subgoal temporal structures combined with theoretically grounded reward shaping provide a practical and scalable solution for long-horizon offline RL.

cs.LG↗

arXiv2026

Coordinated Cooling and Compute Management for AI Datacenters

The AI datacenters are currently being deployed on a large scale to support the training and deployment of power-intensive large-language models (LLMs). Extensive amount of computation and cooling required in datacenters increase concerns about the energy use and carbon emissions of AI datacenters. Although current state-of-the-art has examined the energy efficiency of LLM inference, most prior research focused on optimizing compute-side scheduling without considering thermal objectives or constraints. Since GPU-intensive inference generates substantial heat that can degrade datacenter performance, ignoring thermal effects can increase total energy consumption and reduce the efficiency of LLM serving. To fill this gap, we profile the characteristics of GPU servers under varying cooling and AI jobs, and develop a joint cooling and computing modeling approach for AI datacenters. Built upon such workload and thermal dynamics models, a novel hierarchical control framework is proposed to co-optimize computing and thermal management by identifying the optimal GPU parallelism, frequency (DVFS), and cooling control knobs. Using real Azure inference traces and detailed GPU profiling, our model balances serving latency and thermal constraints in AI datacenters while significantly improving AI datacenters' energy efficiency.

eess.SY↗

arXiv2026

Scaling Laws of Machine Learning for Optimal Power Flow

Optimal power flow (OPF) is one of the fundamental tasks for power system operations. While machine learning (ML) approaches such as deep neural networks (DNNs) have been widely studied to enhance OPF solution speed and performance, their practical deployment faces two critical scaling questions: What is the minimum training data volume required for reliable results? How should ML models' complexity balance accuracy with real-time computational limits? Existing studies evaluate discrete scenarios without quantifying these scaling relationships, leading to trial-and-error-based ML development in real-world applications. This work presents the first systematic scaling study for ML-based OPF across two dimensions: data scale (0.1K-40K training samples) and compute scale (multiple NN architectures with varying FLOPs). Our results reveal consistent power-law relationships on both DNNs and physics-informed NNs (PINNs) between each resource dimension and three core performance metrics: prediction error (MAE), constraint violations and speed. We find that for ACOPF, the accuracy metric scales with dataset size and training compute. These scaling laws enable predictable and principled ML pipeline design for OPF. We further identify the divergence between prediction accuracy and constraint feasibility and characterize the compute-optimal frontier. This work provides quantitative guidance for ML-OPF design and deployments.

cs.LG↗

arXiv2025

FREESH: Fair, Resource- and Energy-Efficient Scheduling for LLM Serving on Heterogeneous GPUs

The ever-increasing computation and energy demand for LLM and AI agents call for holistic and efficient optimization of LLM serving systems. In practice, heterogeneous GPU clusters can be deployed in a geographically distributed manner, while LLM load also observes diversity in terms of both query traffic and serving patterns. LLM queries running on advanced GPUs during a high-emission hour at one location can lead to significantly higher carbon footprints versus same queries running on mid-level GPUs at a low-emission time and location. By observing LLM serving requirements and leveraging spatiotemporal computation flexibility, we consider the joint routing and scheduling problem, and propose FREESH to cooperatively run a group of data centers while minimizing user-specified carbon or energy objectives. FREESH identifies the optimal configurations of balanced load serving by matching distinct GPU instance's power-throughput characteristics with predictable LLM query length and workloads. To ensure both latency and fairness requirements, FREESH identifies optimized parallelism and query routing schedules together with dynamic GPU frequency scaling for power saving, and Least-Laxity-First (LLF) serving strategy for query scheduling. During the 1-hour serving on production workloads, FREESH reduces energy by 28.6% and emissions by 45.45% together with improvements in SLO attainment and fairness.

cs.DC↗

arXiv2025

DeepfakeBench-MM: A Comprehensive Benchmark for Multimodal Deepfake Detection

The misuse of advanced generative AI models has resulted in the widespread proliferation of falsified data, particularly forged human-centric audiovisual content, which poses substantial societal risks (e.g., financial fraud and social instability). In response to this growing threat, several works have preliminarily explored countermeasures. However, the lack of sufficient and diverse training data, along with the absence of a standardized benchmark, hinder deeper exploration. To address this challenge, we first build Mega-MMDF, a large-scale, diverse, and high-quality dataset for multimodal deepfake detection. Specifically, we employ 21 forgery pipelines through the combination of 10 audio forgery methods, 12 visual forgery methods, and 6 audio-driven face reenactment methods. Mega-MMDF currently contains 0.1 million real samples and 1.1 million forged samples, making it one of the largest and most diverse multimodal deepfake datasets, with plans for continuous expansion. Building on it, we present DeepfakeBench-MM, the first unified benchmark for multimodal deepfake detection. It establishes standardized protocols across the entire detection pipeline and serves as a versatile platform for evaluating existing methods as well as exploring novel approaches. DeepfakeBench-MM currently supports 5 datasets and 11 multimodal deepfake detectors. Furthermore, our comprehensive evaluations and in-depth analyses uncover several key findings from multiple perspectives (e.g., augmentation, stacked forgery). We believe that DeepfakeBench-MM, together with our large-scale Mega-MMDF, will serve as foundational infrastructures for advancing multimodal deepfake detection.

cs.CR↗

arXiv2025

Multi-Task Vehicle Routing Solver via Mixture of Specialized Experts under State-Decomposable MDP

Existing neural methods for multi-task vehicle routing problems (VRPs) typically learn unified solvers to handle multiple constraints simultaneously. However, they often underutilize the compositional structure of VRP variants, each derivable from a common set of basis VRP variants. This critical oversight causes unified solvers to miss out the potential benefits of basis solvers, each specialized for a basis VRP variant. To overcome this limitation, we propose a framework that enables unified solvers to perceive the shared-component nature across VRP variants by proactively reusing basis solvers, while mitigating the exponential growth of trained neural solvers. Specifically, we introduce a State-Decomposable MDP (SDMDP) that reformulates VRPs by expressing the state space as the Cartesian product of basis state spaces associated with basis VRP variants. More crucially, this formulation inherently yields the optimal basis policy for each basis VRP variant. Furthermore, a Latent Space-based SDMDP extension is developed by incorporating both the optimal basis policies and a learnable mixture function to enable the policy reuse in the latent space. Under mild assumptions, this extension provably recovers the optimal unified policy of SDMDP through the mixture function that computes the state embedding as a mapping from the basis state embeddings generated by optimal basis policies. For practical implementation, we introduce the Mixture-of-Specialized-Experts Solver (MoSES), which realizes basis policies through specialized Low-Rank Adaptation (LoRA) experts, and implements the mixture function via an adaptive gating mechanism. Extensive experiments conducted across VRP variants showcase the superiority of MoSES over prior methods.

cs.AI↗

arXiv2025

Towards Generalizable PDE Dynamics Forecasting via Physics-Guided Invariant Learning

Advanced deep learning-based approaches have been actively applied to forecast the spatiotemporal physical dynamics governed by partial differential equations (PDEs), which acts as a critical procedure in tackling many science and engineering problems. As real-world physical environments like PDE system parameters are always capricious, how to generalize across unseen out-of-distribution (OOD) forecasting scenarios using limited training data is of great importance. To bridge this barrier, existing methods focus on discovering domain-generalizable representations across various PDE dynamics trajectories. However, their zero-shot OOD generalization capability remains deficient, since extra test-time samples for domain-specific adaptation are still required. This is because the fundamental physical invariance in PDE dynamical systems are yet to be investigated or integrated. To this end, we first explicitly define a two-fold PDE invariance principle, which points out that ingredient operators and their composition relationships remain invariant across different domains and PDE system evolution. Next, to capture this two-fold PDE invariance, we propose a physics-guided invariant learning method termed iMOOE, featuring an Invariance-aligned Mixture Of Operator Expert architecture and a frequency-enriched invariant learning objective. Extensive experiments across simulated benchmarks and real-world applications validate iMOOE's superior in-distribution performance and zero-shot generalization capabilities on diverse OOD forecasting scenarios.

cs.LG↗

arXiv2025

What Deserves Memory: Adaptive Memory Distillation for LLM Agents

Memory systems for LLM agents struggle to determine what information deserves retention. Existing approaches rely on predefined heuristics such as importance scores, emotional tags, or factual templates, encoding designer intuition rather than learning from the data itself. Inspired by cognitive ideas, we propose NEMORI, an adaptive memory distillation framework that casts the assessment of experience's future utility as a matter of predictability. Specifically, NEMORI comprises two cascading modules: Episodic Memory Integration transforms raw interactions into coherent narratives, and Semantic Knowledge Distillation extracts insights via prediction error. Centering on distillation, the framework remains agnostic to downstream management. Extensive experiments confirm that NEMORI achieves strong performance, efficiency, and storage reduction. Our work suggests that observing the intrinsic properties of interaction sequences offers a viable, data-driven alternative to heuristic-based memory design. Code: https://github.com/nemori-ai/nemori.

cs.AI↗

arXiv2025

Carbon Emission Flow Tracing: Fast Algorithm and California Grid Study

Power systems decarbonization are at the focal point of the clean energy transition. While system operators and utility companies increasingly publicize system-level carbon emission information, it remains unclear how emissions from individual generators are transported through the grid and how they impact electricity users at specific locations. This paper presents a novel and computationally efficient approach for exact quantification of nodal average and marginal carbon emission rates, applicable to both AC and DC optimal power flow problems. The approach leverages graph-based topological sorting and directed cycle removal techniques, applied to directed graphs formed by generation dispatch and optimal power flow solutions. Our proposed algorithm efficiently identifies each generator's contribution to each node, capturing how emissions are spatially distributed under varying system conditions. To validate its effectiveness and reveal locational and temporal emission patterns in the real world, we simulate the 8,870-bus realistic California grid using actual CAISO data and the CATS model. Based on year long hourly data on nodal loads and renewable generation, obtained or estimated from CAISO public data, our method accurately estimates power flow conditions, generation mixes, and systemwide emissions, and delivers fine grained spatiotemporal emission analysis for every California county. Both our algorithm and the California study are open-sourced, providing a foundation for future research on grid emissions, planning, operations, and energy policy.

eess.SY↗