arXiv subjects

Zhe Yang

Publications and source records attributed to Zhe Yang.

At least 37 records · Page 2Linked to original sources

arXiv2026

HippoCamp: Benchmarking Contextual Agents on Personal Computers

We present HippoCamp, a new benchmark designed to evaluate agents' capabilities on multimodal file management. Unlike existing agent benchmarks that focus on tasks like web interaction, tool use, or software automation in generic settings, HippoCamp evaluates agents in user-centric environments to model individual user profiles and search massive personal files for context-aware reasoning. Our benchmark instantiates device-scale file systems over real-world profiles spanning diverse modalities, comprising 42.4 GB of data across over 2K real-world files. Building upon the raw files, we construct 581 QA pairs to assess agents' capabilities in search, evidence perception, and multi-step reasoning. To facilitate fine-grained analysis, we provide 46.1K densely annotated structured trajectories for step-wise failure diagnosis. We evaluate a wide range of state-of-the-art multimodal large language models (MLLMs) and agentic methods on HippoCamp. Our comprehensive experiments reveal a significant performance gap: even the most advanced commercial models achieve only 48.3% accuracy in user profiling, struggling particularly with long-horizon retrieval and cross-modal reasoning within dense personal file systems. Furthermore, our step-wise failure diagnosis identifies multimodal perception and evidence grounding as the primary bottlenecks. Ultimately, HippoCamp exposes the critical limitations of current agents in realistic, user-centric environments and provides a robust foundation for developing next-generation personal AI assistants.

cs.AI↗

arXiv2026

GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation

Visual grounding, localizing objects from natural language descriptions, represents a critical bridge between language and vision understanding. While multimodal large language models (MLLMs) achieve impressive scores on existing benchmarks, a fundamental question remains: can MLLMs truly visually ground with human-like sophistication, or are they merely pattern-matching on simplified datasets? Current benchmarks fail to capture real-world complexity where humans effortlessly navigate intricate references and recognize when grounding is impossible. To rigorously assess MLLMs' true capabilities, we introduce GroundingME, a benchmark that systematically challenges models across four critical dimensions: (1) Discriminative: distinguishing highly similar objects, (2) Spatial: understanding complex relational descriptions, (3) Limited: handling occlusions or tiny objects, and (4) Rejection: recognizing ungroundable queries. Through careful curation combining automated generation with human verification, we create 1,005 challenging examples mirroring real-world complexity. Evaluating 25 state-of-the-art MLLMs reveals a profound capability gap: the best model achieves only 45.1% accuracy, while most score 0% on rejection tasks. We explore two strategies for improvements: (1) test-time scaling selects optimal response by thinking trajectory to improve overall performance by up to 4.5%, and (2) data-mixture training boosts rejection accuracy from 0% to 27.9%. GroundingME thus serves as both a diagnostic tool revealing current limitations in MLLMs and a roadmap toward human-level visual grounding. Project page: https://groundingme.github.io

cs.CV↗

arXiv2026

Towards Better RL Training Data Utilization via Second-Order Rollout

Reinforcement Learning (RL) has empowered Large Language Models (LLMs) with strong reasoning capabilities, but vanilla RL mainly focuses on generation capability improvement by training with only first-order rollout (generating multiple responses for a question), and we argue that this approach fails to fully exploit the potential of training data because of the neglect of critique capability training. To tackle this problem, we further introduce the concept of second-order rollout (generating multiple critiques for a response) and propose a unified framework for jointly training generation and critique capabilities. Extensive experiments across various models and datasets demonstrate that our approach can utilize training data more effectively than vanilla RL and achieve better performance under the same training data. Additionally, we uncover several insightful findings regarding second-order rollout and critique training, such as the importance of label balance in critique training and the noise problem of outcome-based rewards, which can be mitigated through sampling techniques. Our work offers a preliminary exploration of dynamic data augmentation and joint generation-critique training in RL, providing meaningful inspiration for the further advancement of RL training

cs.CL↗

arXiv2026

ForensicHub: A Unified Benchmark & Codebase for All-Domain Fake Image Detection and Localization

The field of Fake Image Detection and Localization (FIDL) is highly fragmented, encompassing four domains: deepfake detection (Deepfake), image manipulation detection and localization (IMDL), artificial intelligence-generated image detection (AIGC), and document image manipulation localization (Doc). Although individual benchmarks exist in some domains, a unified benchmark for all domains in FIDL remains blank. The absence of a unified benchmark results in significant domain silos, where each domain independently constructs its datasets, models, and evaluation protocols without interoperability, preventing cross-domain comparisons and hindering the development of the entire FIDL field. To close the domain silo barrier, we propose ForensicHub, the first unified benchmark & codebase for all-domain fake image detection and localization. Considering drastic variations on dataset, model, and evaluation configurations across all domains, as well as the scarcity of open-sourced baseline models and the lack of individual benchmarks in some domains, ForensicHub: i) proposes a modular and configuration-driven architecture that decomposes forensic pipelines into interchangeable components across datasets, transforms, models, and evaluators, allowing flexible composition across all domains; ii) fully implements 10 baseline models, 6 backbones, 2 new benchmarks for AIGC and Doc, and integrates 2 existing benchmarks of DeepfakeBench and IMDLBenCo through an adapter-based design; iii) conducts indepth analysis based on the ForensicHub, offering 8 key actionable insights into FIDL model architecture, dataset characteristics, and evaluation standards. ForensicHub represents a significant leap forward in breaking the domain silos in the FIDL field and inspiring future breakthroughs.

cs.CV↗

arXiv2025

FC-MIR: A Mobile Screen Awareness Framework for Intent-Aware Recommendation based on Frame-Compressed Multimodal Trajectory Reasoning

Identifying user intent from mobile UI operation trajectories is critical for advancing UI understanding and enabling task automation agents. While Multimodal Large Language Models (MLLMs) excel at video understanding tasks, their real-time mobile deployment is constrained by heavy computational costs and inefficient redundant frame processing. To address these issues, we propose the FC-MIR framework: leveraging keyframe sampling and adaptive concatenation, it cuts visual redundancy to boost inference efficiency, while integrating state-of-the-art closed-source MLLMs or fine-tuned models (e.g., Qwen3-VL) for trajectory summarization and intent prediction. We further expand task scope to explore generating post-prediction operations and search suggestions, and introduce a fine-grained metric to evaluate the practical utility of summaries, predictions, and suggestions. For rigorous assessment, we construct a UI trajectory dataset covering scenarios from UI-Agents (Agent-I) and real user interactions (Person-I). Experimental results show our compression method retains performance at 50%-60% compression rates; both closed-source and fine-tuned MLLMs demonstrate strong intent summarization, supporting potential lightweight on-device deployment. However, MLLMs still struggle with useful and "surprising" suggestions, leaving room for improvement. Finally, we deploy the framework in a real-world setting, integrating UI perception and UI-Agent proxies to lay a foundation for future progress in this field.

cs.AI↗

arXiv2025

Stochastic control for Backward Stochastic Differential Equations with semi-Markov chain noises

In this paper, we extend the results of Elliott and Yang \cite{elliott3} and discuss the control of a stochastic process for which the driving noise is provided by a martingale associated with a semi-Markov Chain. An existence and a comparison theorem are obtained. In our discrete time setting, adjoint processes are provided by backward stochastic difference equations. Technical results from partial differential equation theory to establish a verification theorem are not required.

math.PR↗

arXiv2025

Impact and Mitigation of Current Saturation Algorithms in Grid-Forming Inverters on Power Swing Detection

Grid-forming (GFM) inverter-based resources (IBRs) are capable of emulating the external characteristics of synchronous generators (SGs) through the careful design of the control loops. However, the current limiter in the control loops of the GFM IBR poses challenges to the effectiveness of power swing detection functions designed for SG-based systems. Among various current limiting strategies, current saturation algorithms (CSAs) are widely employed for their strict current limiting capability, and are the focus of this paper. The paper presents a theoretical analysis of the conditions for entering and exiting the current saturation mode of the GFM IBR under three CSAs. The corresponding impedance trajectories observed by the relay on the GFM IBR side are investigated. The analysis results reveal that the unique impedance trajectories under these CSAs markedly differ from those associated with SGs. Moreover, it is demonstrated that the conventional power swing detection scheme may lose functionality due to the rapid movement of the trajectory. To mitigate this issue, an optimal current saturation strategy is proposed. Conclusions are validated through simulations in MATLAB/Simulink.

eess.SY↗

arXiv2025

SwitchDelta: Asynchronous Metadata Updating for Distributed Storage with In-Network Data Visibility

Distributed storage systems typically maintain strong consistency between data nodes and metadata nodes by adopting ordered writes: 1) first installing data; 2) then updating metadata to make data visible.We propose SwitchDelta to accelerate ordered writes by moving metadata updates out of the critical path. It buffers in-flight metadata updates in programmable switches to enable data visibility in the network and retain strong consistency. SwitchDelta uses a best-effort data plane design to overcome the resource limitation of switches and designs a novel metadata update protocol to exploit the benefits of in-network data visibility. We evaluate SwitchDelta in three distributed in-memory storage systems: log-structured key-value stores, file systems, and secondary indexes. The evaluation shows that SwitchDelta reduces the latency of write operations by up to 52.4% and boosts the throughput by up to 126.9% under write-heavy workloads.

cs.DC↗

arXiv2025

Enhancing Reliability across Short and Long-Form QA via Reinforcement Learning

While reinforcement learning has unlocked unprecedented complex reasoning in large language models, it has also amplified their propensity for hallucination, creating a critical trade-off between capability and reliability. This work confronts this challenge by introducing a targeted RL framework designed to mitigate both intrinsic and extrinsic hallucinations across short and long-form question answering. We address extrinsic hallucinations (flawed internal knowledge) by creating a novel training set from open-ended conversions of TriviaQA. Concurrently, we tackle intrinsic hallucinations (unfaithfulness to context) by leveraging long-form texts from FineWeb in a fact-grounding reward scheme. To further bolster reliability, our framework explicitly rewards the model for refusing to answer unanswerable questions, thereby cultivating crucial cautiousness. Extensive experiments demonstrate that our methodology yields significant performance gains across a diverse suite of benchmarks, substantially reducing both hallucination types. Ultimately, this research contributes a practical framework for resolving the critical tension between advanced reasoning and factual trustworthiness, paving the way for more capable and reliable large language models.

cs.CL↗

arXiv2025

C$^3$TG: Conflict-aware, Composite, and Collaborative Controlled Text Generation

Recent advancements in large language models (LLMs) have demonstrated remarkable text generation capabilities. However, controlling specific attributes of generated text remains challenging without architectural modifications or extensive fine-tuning. Current methods typically toggle a single, basic attribute but struggle with precise multi-attribute control. In scenarios where attribute requirements conflict, existing methods lack coordination mechanisms, causing interference between desired attributes. Furthermore, these methods fail to incorporate iterative optimization processes in the controlled generation pipeline. To address these limitations, we propose Conflict-aware, Composite, and Collaborative Controlled Text Generation (C$^3$TG), a two-phase framework for fine-grained, multi-dimensional text attribute control. During generation, C$^3$TG selectively pairs the LLM with the required attribute classifiers from the 17 available dimensions and employs weighted KL-divergence to adjust token probabilities. The optimization phase then leverages an energy function combining classifier scores and penalty terms to resolve attribute conflicts through iterative feedback, enabling precise control over multiple dimensions simultaneously while preserving natural text flow. Experiments show that C$^3$TG significantly outperforms baselines across multiple metrics including attribute accuracy, linguistic fluency, and output diversity, while simultaneously reducing toxicity. These results establish C$^3$TG as an effective and flexible solution for multi-dimensional text attribute control that requires no costly model modifications.

cs.CL↗

arXiv2025

Adaptive Redundancy Regulation for Balanced Multimodal Information Refinement

Multimodal learning aims to improve performance by leveraging data from multiple sources. During joint multimodal training, due to modality bias, the advantaged modality often dominates backpropagation, leading to imbalanced optimization. Existing methods still face two problems: First, the long-term dominance of the dominant modality weakens representation-output coupling in the late stages of training, resulting in the accumulation of redundant information. Second, previous methods often directly and uniformly adjust the gradients of the advantaged modality, ignoring the semantics and directionality between modalities. To address these limitations, we propose Adaptive Redundancy Regulation for Balanced Multimodal Information Refinement (RedReg), which is inspired by information bottleneck principle. Specifically, we construct a redundancy phase monitor that uses a joint criterion of effective gain growth rate and redundancy to trigger intervention only when redundancy is high. Furthermore, we design a co-information gating mechanism to estimate the contribution of the current dominant modality based on cross-modal semantics. When the task primarily relies on a single modality, the suppression term is automatically disabled to preserve modality-specific information. Finally, we project the gradient of the dominant modality onto the orthogonal complement of the joint multimodal gradient subspace and suppress the gradient according to redundancy. Experiments show that our method demonstrates superiority among current major methods in most scenarios. Ablation experiments verify the effectiveness of our method. The code is available at https://github.com/xia-zhe/RedReg.git

cs.LG↗

arXiv2025

ZJUNlict Extended Team Description Paper 2025

This paper presents the ZJUNlict team's work over the past year, covering both hardware and software advancements. In the hardware domain, the integration of an IMU into the v2023 robot was completed to enhance posture accuracy and angular velocity planning. On the software side, key modules were optimized, including the strategy and CUDA modules, with significant improvements in decision making efficiency, ball pursuit prediction, and ball possession prediction to adapt to high-tempo game dynamics.

cs.RO↗

arXiv2025

Towards an AI Musician: Synthesizing Sheet Music Problems for Musical Reasoning

Enhancing the ability of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) to interpret sheet music is a crucial step toward building AI musicians. However, current research lacks both evaluation benchmarks and training data for sheet music reasoning. Inspired by mathematics, where simple operations yield infinite verifiable problems, we introduce a novel approach that treats core music theory rules, such as those governing beats and intervals, as programmatic functions to systematically synthesize a vast and diverse corpus of sheet music reasoning problems. This approach allows us to introduce a data synthesis framework that generates verifiable sheet music questions in both textual and visual modalities, leading to the Synthetic Sheet Music Reasoning Benchmark (SSMR-Bench) and a complementary training set. Evaluation results on SSMR-Bench highlight the key role reasoning plays in interpreting sheet music, while also pointing out the ongoing challenges in understanding sheet music in a visual format. By leveraging synthetic data for RLVR, all models show significant improvements on the SSMR-Bench. Additionally, they also demonstrate considerable advancements on previously established human-crafted benchmarks, such as MusicTheoryBench and the music subset of MMMU. Finally, our results show that the enhanced reasoning ability can also facilitate music composition.

cs.CL↗

arXiv2025

A Probabilistic Inference Scaling Theory for LLM Self-Correction

Large Language Models (LLMs) have demonstrated the capability to refine their generated answers through self-correction, enabling continuous performance improvement over multiple rounds. However, the mechanisms underlying how and why accuracy evolves during this iterative process remain unexplored. To fill this gap, we propose a probabilistic theory to model the dynamics of accuracy change and explain the performance improvements observed in multi-round self-correction. Through mathematical derivation, we establish that the accuracy after the $t^{th}$ round of self-correction is given by: $Acc_t = Upp - α^t(Upp - Acc_0),$ where $Acc_0$ denotes the initial accuracy, $Upp$ represents the upper bound of accuracy convergence, and $α$ determines the rate of convergence. Based on our theory, these parameters can be calculated and the predicted accuracy curve then can be obtained through only a single round of self-correction. Extensive experiments across diverse models and datasets demonstrate that our theoretical predictions align closely with empirical accuracy curves, validating the effectiveness of the theory. Our work provides a theoretical foundation for understanding LLM self-correction, thus paving the way for further explorations.

cs.CL↗

arXiv2025

Capturing quantum phase transition in the ultraviolet region by holography

We reveal for the first time that ultraviolet (UV) observables can diagnose quantum phase transitions (QPTs). In a class of holographic models exhibiting metal-insulator transitions, we study two types of UV observables -- high-frequency conductivity and short-range entanglement. Remarkably, we find that the derivatives of these UV observables exhibit extrema near the quantum critical point. Analytical results show these critical behaviors arise from the deformation of the asymptotic bulk geometry. Moreover, these UV diagnostics show enhanced robustness to thermal fluctuations compared to typical infrared (IR) diagnostics, providing a clean method to identify quantum criticality at finite temperature. This work opens a new window for exploring quantum critical phenomena via UV physics in the laboratory.

hep-th↗

arXiv2025

UMRE: A Unified Monotonic Transformation for Ranking Ensemble in Recommender Systems

Industrial recommender systems commonly rely on ensemble sorting (ES) to combine predictions from multiple behavioral objectives. Traditionally, this process depends on manually designed nonlinear transformations (e.g., polynomial or exponential functions) and hand-tuned fusion weights to balance competing goals -- an approach that is labor-intensive and frequently suboptimal in achieving Pareto efficiency. In this paper, we propose a novel Unified Monotonic Ranking Ensemble (UMRE) framework to address the limitations of traditional methods in ensemble sorting. UMRE replaces handcrafted transformations with Unconstrained Monotonic Neural Networks (UMNN), which learn expressive, strictly monotonic functions through the integration of positive neural integrals. Subsequently, a lightweight ranking model is employed to fuse the prediction scores, assigning personalized weights to each prediction objective. To balance competing goals, we further introduce a Pareto optimality strategy that adaptively coordinates task weights during training. UMRE eliminates manual tuning, maintains ranking consistency, and achieves fine-grained personalization. Experimental results on two public recommendation datasets (Kuairand and Tenrec) and online A/B tests demonstrate impressive performance and generalization capabilities.

cs.IR↗

arXiv2025

Power Swing Trajectory Influenced by Virtual Impedance-Based Current-Limiting Strategy

Grid-forming (GFM) inverter-based resources (IBRs) can emulate the external characteristics of synchronous generators (SGs) through appropriate control loop design. However, in systems with GFM IBRs, the apparent impedance trajectory under current limitation differs significantly from that of SG-based systems due to the limited overcurrent capability of power electronic devices. This difference challenges the power swing detection functions of distance relays designed for SG-based systems. This paper presents a theoretical analysis of the apparent impedance trajectory over a full power swing cycle under two typical current-limiting strategies: variable virtual impedance (VI) and adaptive VI. The analysis reveals that the trajectory under VI current-limiting strategies differs significantly from that of a conventional SG. The results also indicate that the control parameters affect the characteristics of the trajectory. In addition, the new trajectories challenge conventional power swing detection functions, increasing the risk of malfunction. Furthermore, the implementation of VI leads to a deterioration in system stability. The theoretical analysis is further validated through simulations on the MATLAB/Simulink platform.

eess.SY↗

arXiv2025

Palette of Language Models: A Solver for Controlled Text Generation

Recent advancements in large language models have revolutionized text generation with their remarkable capabilities. These models can produce controlled texts that closely adhere to specific requirements when prompted appropriately. However, designing an optimal prompt to control multiple attributes simultaneously can be challenging. A common approach is to linearly combine single-attribute models, but this strategy often overlooks attribute overlaps and can lead to conflicts. Therefore, we propose a novel combination strategy inspired by the Law of Total Probability and Conditional Mutual Information Minimization on generative language models. This method has been adapted for single-attribute control scenario and is termed the Palette of Language Models due to its theoretical linkage between attribute strength and generation style, akin to blending colors on an artist's palette. Moreover, positive correlation and attribute enhancement are advanced as theoretical properties to guide a rational combination strategy design. We conduct experiments on both single control and multiple control settings, and achieve surpassing results.

cs.CL↗