arXiv subjects

Haotian Zhao

Publications and source records attributed to Haotian Zhao.

15 recordsLinked to original sources

arXiv2026

WebGameBench: Requirement-to-Application Evaluation for Coding Agents via Browser-Native Games

Coding agents are increasingly used as application builders, yet many evaluations still focus on source code, repository-level tests, or intermediate traces rather than the delivered application. We introduce WebGameBench, a requirement-to-application benchmark that evaluates whether coding agents can turn a frozen Structured WebGame Specification into a browser-accessible game. Browser-native games provide a compact but behavior-dense testbed: even simple games require coordinated input handling, spatial mapping, rule execution, state transitions, terminal conditions, restart behavior, and visible feedback. In WebGameBench, each generated artifact is built, served, and exposed as a browser-accessible application under a unified deployment protocol. A runtime evaluator then interacts with the delivered game in a real browser and assigns a three-way label: EXCELLENT, USABLE, or UNUSABLE. On a human-reviewed subset, the runtime label is broadly aligned with human gameplay review under the Usable-rate criterion. Across 111 tasks, 12 coding agents, and 14 evaluation configurations, WebGameBench separates current systems: the best configuration reaches a 76.9% usable rate but only a 20.2% excellent rate. This gap shows that crossing the minimum playable-delivery threshold is still far from complete requirement satisfaction. To our knowledge, WebGameBench is the first requirement-to-application benchmark for browser-native game delivery that validates delivered-application runtime labels against independent human gameplay review under the Usable-rate criterion.

cs.AI↗

arXiv2026

Rollout Pass-Rate Control: Steering Binary-Reward RL Toward Its Most Informative Regime

Agentic reinforcement learning (RL) for software engineering spends much of its compute on stateful trajectories whose grouped binary rewards are highly skewed and weakly contrastive. We frame this as pass-rate control and show that the binary reward-side signal is strongest near a 50% rollout pass rate under four criteria: reward entropy, group-filtering survival, leave-one-out (RLOO) advantage energy under Group Relative Policy Optimization (GRPO), and success-failure pair count. We propose Prefix Sampling (PS), which replays self-generated trajectory prefixes to steer skewed groups toward this regime: successful prefixes give mostly failing groups a head start, while failing prefixes handicap mostly passing groups. Replayed states are reconstructed through the existing rollout path, and replayed tokens are masked from the loss so optimization applies only to current-policy continuations. On SWE-bench Verified, PS reaches the baseline high-score regime within evaluation variability while delivering 2.01x and 1.55x end-to-end wall-clock speedups on Qwen3-14B and Qwen3-32B; the 14B peak improves from 0.274 to 0.295. AIME 2025 experiments on 4B and 8B show the same pass-rate-control pattern, and 4B ablations attribute gains to replay, bidirectional coverage, and adaptive control.

cs.LG↗

arXiv2026

AEM: Adaptive Entropy Modulation for Multi-Turn Agentic Reinforcement Learning

Reinforcement learning (RL) has substantially improved the ability of large language model (LLM) agents to interact with environments and solve multi-turn tasks. However, effective agentic RL remains challenging: sparse outcome-only rewards provide limited guidance for assigning credit to individual steps within long interaction trajectories. Existing approaches often introduce dense intermediate supervision, such as process reward models or auxiliary self-supervised signals, which increases supervision and tuning complexity and may limit generalization across tasks and domains. We present AEM, a supervision-free credit assignment method that adaptively modulates entropy dynamics during RL training to improve the exploration-exploitation trade-off. Since in agentic RL the environment is typically affected by a complete response, rather than an individual token, our analysis lifts entropy dynamics from the token level to the response level, aligning uncertainty estimation with the effective action granularity of LLM agents and reducing sensitivity to token-level sampling noise. We further show that entropy drift under natural-gradient updates is governed by the interaction between the sampled-response advantage and its relative surprisal. Motivated by this result, AEM derives a practical response-level uncertainty proxy and uses it to rescale advantages, leveraging the evolving balance between positive and negative samples to naturally transition from exploration to exploitation. Extensive experiments on ALFWorld, WebShop, and SWE-bench-Verified with models ranging from 1.5B to 32B demonstrate that AEM consistently improves strong RL baselines, including a +1.4\% gain when integrated into a state-of-the-art software-engineering RL training framework.

cs.AI↗

arXiv2026

A Unified Heterogeneous Implementation of Numerical Atomic Orbitals-Based Real-Time TDDFT within the ABACUS Package

We present a unified heterogeneous computing framework for real-time time-dependent density functional theory (RT-TDDFT) based on numerical atomic orbitals (NAOs), implemented in the ABACUS package. We introduce three co-designed abstraction layers, including unified data containers, unified linear algebra operators, and unified grid integration interfaces. These layers collectively accelerate the two most demanding parts of NAO-based RT-TDDFT: explicit real-time wavefunction propagation and real-space grid operations such as Hamiltonian construction and force evaluation under external fields. We validate the method by computing optical properties for systems ranging from finite molecules to periodic solids, showing excellent agreement with standard benchmarks. Performance evaluations on bulk silicon demonstrate that a single GPU can achieve substantial wall-clock speedup over a fully utilized dual-socket CPU node. Furthermore, distributed multi-GPU strong-scaling tests confirm high parallel efficiency over tens of GPUs. This work establishes a high-performance, portable platform for large-scale first-principles simulations of ultrafast electron dynamics.

cond-mat.mtrl-sci↗

arXiv2026

LOGIGEN: Logic-Driven Generation of Verifiable Agentic Tasks

The evolution of Large Language Models (LLMs) from static instruction-followers to autonomous agents necessitates operating within complex, stateful environments to achieve precise state-transition objectives. However, this paradigm is bottlenecked by data scarcity, as existing tool-centric reverse-synthesis pipelines fail to capture the rigorous logic of real-world applications. We introduce \textbf{LOGIGEN}, a logic-driven framework that synthesizes verifiable training data based on three core pillars: \textbf{Hard-Compiled Policy Grounding}, \textbf{Logic-Driven Forward Synthesis}, and \textbf{Deterministic State Verification}. Specifically, a Triple-Agent Orchestration is employed: the \textbf{Architect} compiles natural-language policy into database constraints to enforce hard rules; the \textbf{Set Designer} initializes boundary-adjacent states to trigger critical policy conflicts; and the \textbf{Explorer} searches this environment to discover causal solution paths. This framework yields a dataset of 20,000 complex tasks across 8 domains, where validity is strictly guaranteed by checking exact state equivalence. Furthermore, we propose a verification-based training protocol where Supervised Fine-Tuning (SFT) on verifiable trajectories establishes compliance with hard-compiled policy, while Reinforcement Learning (RL) guided by dense state-rewards refines long-horizon goal achievement. On $\tau^2$-Bench, LOGIGEN-32B(RL) achieves a \textbf{79.5\% success rate}, substantially outperforming the base model (40.7\%). These results demonstrate that logic-driven synthesis combined with verification-based training effectively constructs the causally valid trajectories needed for next-generation agents.

cs.AI↗

arXiv2026

SWE-Hub: A Unified Production System for Scalable, Executable Software Engineering Tasks

Progress in software-engineering agents is increasingly constrained by the scarcity of executable, scalable, and realistic data for training and evaluation. This scarcity stems from three fundamental challenges in existing pipelines: environments are brittle and difficult to reproduce across languages; synthesizing realistic, system-level bugs at scale is computationally expensive; and existing data predominantly consists of short-horizon repairs, failing to capture long-horizon competencies like architectural consistency. We introduce \textbf{SWE-Hub}, an end-to-end system that operationalizes the data factory abstraction by unifying environment automation, scalable synthesis, and diverse task generation into a coherent production stack. At its foundation, the \textbf{Env Agent} establishes a shared execution substrate by automatically converting raw repository snapshots into reproducible, multi-language container environments with standardized interfaces. Built upon this substrate, \textbf{SWE-Scale} engine addresses the need for high-throughput generation, combining cross-language code analysis with cluster-scale validation to synthesize massive volumes of localized bug-fix instances. \textbf{Bug Agent} generates high-fidelity repair tasks by synthesizing system-level regressions involving cross-module dependencies, paired with user-like issue reports that describe observable symptoms rather than root causes. Finally, \textbf{SWE-Architect} expands the task scope from repair to creation by translating natural-language requirements into repository-scale build-a-repo tasks. By integrating these components, SWE-Hub establishes a unified production pipeline capable of continuously delivering executable tasks across the entire software engineering lifecycle.

cs.AI↗

arXiv2025

Real-time time-dependent density functional theory simulations with range-separated hybrid functionals for periodic systems

Real-time time-dependent density functional theory (RT-TDDFT) is a powerful approach for investigating various ultrafast phenomena in materials. However, most existing RT-TDDFT studies rely on adiabatic local or semi-local approximations, which suffer from several shortcomings, including the inability to accurately capture excitonic effects in periodic systems. Combining RT-TDDFT with range-separated hybrid (RSH) functionals has emerged as an effective strategy to overcome these limitations. The RT-TDDFT-RSH implementation for periodic systems requires careful treatment of the Coulomb singularity and choosing proper gauges for the incorporation of external fields. We benchmark two schemes for treating the Coulomb singularity - the truncated Coulomb potential and the auxiliary-function correction - and find that the latter shows better convergence behavior and numerical stability for long-range corrected hybrid functions. Additionally, we assess the impact of gauge choice in simulations using numerical atomic orbitals and show that the recently proposed hybrid gauge incorporating position-dependent phases provides a more accurate description of excitonic absorption than the conventional velocity gauge. Our implementation significantly improves the accuracy of RT-TDDFT-RSH for modeling ultrafast excitonic dynamics in periodic systems.

cond-mat.mtrl-sci↗

arXiv2025

FingerTip 20K: A Benchmark for Proactive and Personalized Mobile LLM Agents

Mobile GUI agents are becoming critical tools to improve user experience on smart devices, with multimodal large language models (MLLMs) emerging as the dominant paradigms in this domain. Current agents, however, rely on explicit human instructions, overlooking the potential to leverage the contextual information (like location, time, user profile) and historical data for proactive task suggestions. Besides, previous works focus on optimizing the success rate during task execution, but pay less attention to the personalized execution trajectory, thereby neglecting potentially vast differences in user preferences. To address these challenges, we introduce the FingerTip 20K benchmark. We collected 20K unique human demonstrations of multi-step Android device interactions across a variety of everyday apps. These demonstrations are not isolated but are continuously acquired from the users' long-term usage in their real lives, and encompass essential user-related contextual information. The benchmark contains two new tracks: proactive task suggestions by analyzing environment observation and users' previous intents, and personalized task execution by catering to users' action preferences. Our experiments reveal that the tracks we propose pose significant challenges for leveraging user-related information in GUI tasks. We also performed a human study to show that there exists a huge gap between existing agents and humans. The model fine-tuned with the data we collected effectively utilized user information and achieved good results, highlighting the potential of our approach in building more user-oriented mobile LLM agents. Our code is open-source at https://github.com/tsinghua-fib-lab/FingerTip-20K for reproducibility.

cs.HC↗

arXiv2025

Arithmetic properties of Cantor sets involving non-diagonal forms

We show conditions on $k$ such that any number $x$ in the interval $[0, k/2]$ can be represented in the form $x_1^{a_1} x_2^{a_2} + x_3^{a_3} x_4^{a_4} + \cdots + x_{k-1}^{a_{k-1}} x_k^{a_k}$, where the exponents $a_{2i-1}$ and $a_{2i}$ are positive integers satisfying $a_{2i-1} + a_{2i} = s$ for $i = 1, 2, \dots, k/2$, and each $x_i$ belongs to the generalized Cantor set. Moreover, we discuss different types of non-diagonal polynomials and clarify the optimal results in low-dimensional cases.

math.NT↗

arXiv2025

Kimi k1.5: Scaling Reinforcement Learning with LLMs

Language model pretraining with next token prediction has proved effective for scaling compute but is limited to the amount of available training data. Scaling reinforcement learning (RL) unlocks a new axis for the continued improvement of artificial intelligence, with the promise that large language models (LLMs) can scale their training data by learning to explore with rewards. However, prior published work has not produced competitive results. In light of this, we report on the training practice of Kimi k1.5, our latest multi-modal LLM trained with RL, including its RL training techniques, multi-modal data recipes, and infrastructure optimization. Long context scaling and improved policy optimization methods are key ingredients of our approach, which establishes a simplistic, effective RL framework without relying on more complex techniques such as Monte Carlo tree search, value functions, and process reward models. Notably, our system achieves state-of-the-art reasoning performance across multiple benchmarks and modalities -- e.g., 77.5 on AIME, 96.2 on MATH 500, 94-th percentile on Codeforces, 74.9 on MathVista -- matching OpenAI's o1. Moreover, we present effective long2short methods that use long-CoT techniques to improve short-CoT models, yielding state-of-the-art short-CoT reasoning results -- e.g., 60.8 on AIME, 94.6 on MATH500, 47.3 on LiveCodeBench -- outperforming existing short-CoT models such as GPT-4o and Claude Sonnet 3.5 by a large margin (up to +550%).

cs.AI↗

arXiv2025

Risk Assessment and Vulnerability Identification of Energy-Transportation Infrastructure Systems to Extreme Weather

The interaction between extreme weather events and interdependent critical infrastructure systems involves complex spatiotemporal dynamics. Multi-type emergency decisions within energy-transportation infrastructures significantly influence system performance throughout the extreme weather process. A comprehensive assessment of these factors faces challenges in model complexity, heterogeneous differences between energy and transportation systems, and cross-sector privacy. This paper proposes a risk assessment framework that integrates the heterogeneous energy and transportation systems in the form of a unified network flow model, which enables full accommodation of multiple types of energy-transportation emergency decisions while capturing the compound spatiotemporal impacts of extreme weather on both systems simultaneously. Based on this framework, a targeted method for identifying system vulnerabilities is further developed. This method employs neural network surrogates to achieve privacy protection and accelerated identification while maintaining consideration of system interdependencies. Numerical experiments demonstrate that the proposed framework and method can reveal the risk levels faced by urban infrastructure systems, identify vulnerabilities that should be prioritized for reinforcement, and strike a balance between accuracy and speed.

eess.SY↗

arXiv2025

Hybrid Gauge Approach for Accurate Real-Time TDDFT Simulations with Numerical Atomic Orbitals

Ultrafast real-time dynamics are critical for understanding a broad range of physical processes. Real-time time-dependent density functional theory (rt-TDDFT) has emerged as a powerful computational tool for simulating these dynamics, offering insight into untrafast processes and light-matter interactions. In periodic systems, the velocity gauge is essential because it preserves the system's periodicity under an external electric field. Numerical atomic orbitals (NAOs) are widely employed in rt-TDDFT codes due to their efficiency and localized nature. However, directly applying the velocity gauge within the NAO basis set neglects the position-dependent phase variations within atomic orbitals induced by the vector potential, leading to significant computational errors-particularly in current calculations. To resolve this issue, we develop a hybrid gauge that incorporates both the electric field and the vector potential, preserving the essential phase information in atomic orbitals and thereby eliminating these errors. Our benchmark results demonstrate that the hybrid gauge fully resolves the issues encountered with the velocity gauge in NAO-based calculations, providing accurate and reliable results. This algorithm offers a robust framework for future studies on ultrafast dynamics in periodic systems using NAO bases.

cond-mat.mtrl-sci↗

arXiv2024

Reconstructing Quantitative Cerebral Perfusion Images Directly From Measured Sinogram Data Acquired Using C-arm Cone-Beam CT

To shorten the door-to-puncture time for better treating patients with acute ischemic stroke, it is highly desired to obtain quantitative cerebral perfusion images using C-arm cone-beam computed tomography (CBCT) equipped in the interventional suite. However, limited by the slow gantry rotation speed, the temporal resolution and temporal sampling density of typical C-arm CBCT are much poorer than those of multi-detector-row CT in the diagnostic imaging suite. The current quantitative perfusion imaging includes two cascaded steps: time-resolved image reconstruction and perfusion parametric estimation. For time-resolved image reconstruction, the technical challenge imposed by poor temporal resolution and poor sampling density causes inaccurate quantification of the temporal variation of cerebral artery and tissue attenuation values. For perfusion parametric estimation, it remains a technical challenge to appropriately design the handcrafted regularization for better solving the associated deconvolution problem. These two challenges together prevent obtaining quantitatively accurate perfusion images using C-arm CBCT. The purpose of this work is to simultaneously address these two challenges by combining the two cascaded steps into a single joint optimization problem and reconstructing quantitative perfusion images directly from the measured sinogram data. In the developed direct cerebral perfusion parametric image reconstruction technique, TRAINER in short, the quantitative perfusion images have been represented as a subject-specific conditional generative model trained under the constraint of the time-resolved CT forward model, perfusion convolutional model, and the subject's own measured sinogram data. Results shown in this paper demonstrated that using TRAINER, quantitative cerebral perfusion images can be accurately obtained using C-arm CBCT in the interventional suite.

eess.IV↗

arXiv2024

Active Support of Inverters for Improving Short-Term Voltage Security in 100% IBRsPenetrated Power Systems

Due to the energy crisis and environmental pollution, the installed capacity of inverter-based resources (IBRs) in power grids is rapidly increasing, and grid-following control (GFL) is the most prevalent at present. Meanwhile, grid-forming control-based (GFM) devices have been installed in the grid to provide active support for frequency and voltage. In the future GFL devices combined with GFM will be promising, especially in power systems with high penetration or 100% IBRs. When a short-circuit fault occurs in the grid, the controlled current source characteristic of the GFL devices leads to insufficient dynamic voltage support (DVS), while the GFM devices usually reduce the internal voltage to limit the current. Thus, deep voltage sags and undesired disconnections of IBRs may occur. Moreover, due to the dispersed locations and the control strategies' diversity of IBRs, the voltage support of different devices may not be fully coordinated, which is not conducive to short-term voltage security (STVS). To address this issue, a control scheme based on the simulation of transient characteristics of synchronous machines (SMs) is proposed. Then, a new fault ride-through strategy (FRT) is proposed based on the characteristic differences between GFL and GFM devices, and an optimization model of multi-device control parameters is formulated to meet the short-term voltage security constraints (SVSCs) and device capacity constraints. Finally, a fast solution method based on analytical modeling is proposed for the model. Test results based on the doublegenerator-one-load system, the IEEE 14-bus system, and other systems of different sizes show that the proposed method can effectively enhance the active support capability of GFL and GFM to the grid voltage, and avoid the large-scale disconnection of IBRs

eess.SY↗

arXiv2023

Short-Term Voltage Security Constrained UC to Prevent Trip Faults in High Wind Power Penetrated Power Systems

For high wind power-penetrated power systems, the multiple renewable energy station short-circuit ratio (MRSCR) is often insufficient due to weak grid structures. Additionally, transient voltage sag/overvoltage issues may cause trip faults of wind turbines (WTs). Due to the time delay in WTs' controllers, it is difficult for WTs alone to meet the reactive power demands in different stages of the transient process. Some synchronous machines (SMs) must be retained through unit commitment (UC) scheduling to improve MRSCR and prevent trip faults of WTs. The MRSCR and short-term voltage security constrained-UC model is a mixed integer nonlinear programming (MINLP) problem with differential algebraic equations (DAEs) and symbolic matrix inversion, which is intractable to solve. Based on the dynamic characteristics of different devices, the original model is simplified as a general MINLP model without DAEs. Then, generalized Benders decomposition is applied to improve the solution efficiency. The relaxed MRSCR constraints are formulated in the master problem to improve the convergence, and the precise MRSCR constraints are formulated in the subproblems to consider the impact of voltage profiles. Case studies based on several benchmark systems and a provincial power grid verify the validity and efficiency of the proposed method

eess.SY↗