arXiv subjects

Yi Chen

Publications and source records attributed to Yi Chen.

At least 19 recordsLinked to original sources

arXiv2026

Distance Is Not Enough: Forget-Retain Alignment Gap Predicts LLM Relearning Robustness

Machine unlearning aims to make a model forget specific data, yet unlearned LLMs often fail to stay unlearned: brief fine-tuning can revive removed knowledge. Existing robustness predictors rely on global weight-space displacement, but distance alone can be misleading when random or destructive updates collapse performance. We argue that relearning robustness depends on update structure: robust unlearning should affect forget-critical weights while sparing retain-critical ones. We introduce the Forget-Retain Alignment Gap (FRAG), a training-free predictor that scores an update's forget-retain alignment without running a relearning attack, and separates selective from dense updates more reliably than global distance. Building on the forget-critical, retain-sparing principle, Forget-Retain Pruning (FRP) improves relearning robustness. Our results suggest that weight selectivity better explains robustness than distance alone. Code is available at https://github.com/Yi1-Chen/FRAG.

cs.AI↗

arXiv2026

Algorithmic Attention and Content Creation on Social Media Platforms

We study revenue-maximizing attention allocation on an ad-funded social media platform governed by recommendation algorithms. Attention is costly and can be monetized through advertising or allocated to increase creators' exposure, creating a trade-off between monetization and production incentives. In a two-sided model with heterogeneous viewers and creators under private information, the optimal recommendation mix includes content that is ex post suboptimal for some viewers to leverage network externalities. These distortions are targeted: low-ability creators are excluded, while high-ability creators are subsidized through exposure or monetary payments. Two-sided complementarities reshape content variety and quality, with implications for personalization regulation and advertising markets.

econ.TH↗

arXiv2026

Role-aware Heuristic Episodic Attention for Conversational LLMs

Large language models often lose track of persistent instructions and relevant information as multi-turn conversations grow. We study this cumulative contextual decay through three related failure modes: attention pollution, dilution, and drift. We propose REA(Role-aware Heuristic Episodic Attention), a context-management framework that assigns different persistence and representation policies to instructions and episodic interactions. Instructional Memory retains identified global constraints in a dedicated prefix. Episodic Memory preserves user inputs and compresses model replies, while heuristic retrieval selects raw text, compressed representations, or omission for each historical turn. On Long-MT-Bench+, REA improves the judge score from 6.32 to 7.36 on a 10-point scale, a 16.5% relative gain over the Vanilla baseline, and reduces average latency by 2.91x. Additional evaluations show aggregate gains on three backbones spanning 1.7B-7B parameters and on Chinese and English role-playing tasks. These results support role-aware context management as a practical approach to maintaining conversational continuity and instruction adherence.

cs.CL↗

arXiv2026

When Core-Network Elements Become Reachable: Field-Level Exploitability Analysis in Diameter

The increasing exposure of cellular core-network elements to public networks has turned Diameter signaling interfaces into a practical attack surface. While large language models (LLMs) show promise for protocol exploitability analysis, we find that their judgments on Diameter can be unreliable because they may deviate from the tacit analysis discipline followed by security experts and reach conclusions without considering complete protocol evidence. To address this challenge, we present LENS, an LLM-based system for reliable field-level Diameter exploitability analysis. LENS guides LLMs using both exploitability-analysis discipline and the protocol evidence needed for analysis. It infers analysis discipline from existing exploit writeups and evidence-collection guidance from exploitability-analysis failures, enabling systematic evidence collection, exploitability assessment, and proof-of-concept (PoC) attack generation when applicable. Our experiments show that LENS performs reliable analysis across four LLMs and repeated runs, achieving mean recall of 94.00% to 98.00%, precision of 78.73% to 88.94%, and accuracy of 84.00% to 92.00%. We find that more than 25% of field paths in the S6a and Cx interfaces are exploitable under the considered threat model. We further assess their impact on practical implementations and uncover 95 previously unknown vulnerabilities, including 17 in open-source EPC/IMS implementations and 78 in commercial deployments, causing service disruption, information leakage, subscriber tracking, and accelerated battery depletion. Applying LENS to 16 additional protocols spanning cellular, network, industrial, and host-internal systems uncovers 118 previously unknown vulnerabilities, demonstrating its strong generalizability.

cs.CR↗

arXiv2026

Promotion Signaling and Signal Jamming: An Analysis of Job Title Inflation

Starting with Waldman (1984), it has become widely understood that an important role of promotions in labor markets is that promotions serve as signals of high worker ability. In this paper we explore the common and growing practice called job title inflation which serves a signal jamming role that reduces the signaling value of promotions. We first explore a two-period model and show how job title inflation can arise in equilibrium and increase social welfare both because workers directly value impressive sounding job titles and because the practice reduces job assignment inefficiencies due to signaling. We then introduce into the model an investment choice for young workers concerning human capital accumulation and show that job title inflation can decrease rather than increase welfare in that case. That is, when promotions and the associated wage increases serve as an important incentive for investing in human capital, the use of job title inflation can persist even though the practice lowers both overall efficiency and expected worker utility. We then consider three extensions: (i) a setting characterized by worker turnover; (ii) a setting characterized by a wage transparency law; and (iii) a setting characterized by heterogeneous worker visibility. Also, in addition to showing how job title inflation can arise as an equilibrium phenomenon that can either raise or lower welfare, we discuss the extent to which our theoretical analysis matches evidence concerning the use of job title inflation in real-world labor markets.

econ.TH↗

arXiv2026

Agentic AI -- Physicist Collaboration in Experimental Particle Physics: A Proof-of-Concept Measurement with LEP Open Data

We present an agentic measurement of the thrust distribution in $e^{+}e^{-}$ collisions at $\sqrt{s}=91.2$ GeV using ALEPH archived data. This work was carried out through an AI agent-physicist collaboration, using interactive natural-language prompting of Anthropic Claude and OpenAI Codex. Under physicist guidance, the agents wrote and executed all analysis code and drafted the majority of the text included in this paper. A fully-corrected thrust spectrum is obtained via Iterative Bayesian Unfolding and Monte Carlo based corrections. The principal contribution of this work is the demonstration of the agentic analysis, with thrust serving as a benchmark observable for the method. Dedicated precision measurements of thrust are reported in the community elsewhere. This work represents a step toward a theory-experiment loop in which AI agents assist with experimental measurements and theoretical calculations, and synthesize insights by comparing the results, thereby accelerating the cycle that drives discovery in fundamental physics. Our work suggests that precision physics, leveraging the open LEP data and advanced theoretical landscape, provides an ideal testing ground for developing AI agent systems for scientific applications.

hep-ex↗

arXiv2026

AIPOCH Open-Science: An Open-Source AI Research Workbench with Reproducible .science Packages

AI-assisted research leaves instructions, computations, artifact revisions and checks whose relationships matter when another researcher receives the work. We present AIPOCH Open-Science, an open-source research workbench with portable .science session packages, and specify an exchange contract for these recorded relationships. Version and assessment targets remain attributable while receiving-workspace identifiers can be remapped, eligible contents can be omitted explicitly, and imported history remains separate from a writable fork. A controlled session tested an incorrect first version, a reviewed correction, an unreviewed revision and a planned but unexecuted step. We compared complete and partial exports with source records extracted independently of the exporter, then inspected imports into separate projects within the same installation. Both conditions retained the tested version and review associations, kept the unreviewed revision distinct from earlier assessments and represented excluded content as unavailable. A documentation-rich ordinary directory also supplied evidence for all seven static inspection questions. A separate registry case connected a table to its recorded task, execution and review; offline execution of its retained script and fixed input reproduced two saved CSV tables byte for byte. The study contributes an implemented exchange contract and a controlled examination of its preservation rules, content-selection constraints and interpretation boundaries.

cs.SE↗

arXiv2026

UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling

Scaling humanoid foundation models is bottlenecked by the scarcity of robotic data. While massive egocentric human data offers a scalable alternative, bridging the cross-embodiment chasm remains a fundamental challenge due to kinematic mismatches. We introduce UniT (Unified Latent Action Tokenizer via Visual Anchoring), a framework that establishes a unified physical language for human-to-humanoid transfer. Grounded in the philosophy that heterogeneous kinematics share universal visual consequences, UniT employs a tri-branch cross-reconstruction mechanism: actions predict vision to anchor kinematics to physical outcomes, while vision reconstructs actions to filter out irrelevant visual confounders. Concurrently, a fusion branch synergies these purified modalities into a shared discrete latent space of embodiment-agnostic physical intents. We validate UniT across two paradigms: 1) Policy Learning (VLA-UniT): By predicting these unified tokens, it effectively leverages diverse human data to achieve state-of-the-art data efficiency and robust out-of-distribution (OOD) generalization on both humanoid simulation benchmark and real-world deployments, notably demonstrating zero-shot task transfer. 2) World Modeling (WM-UniT): By aligning cross-embodiment dynamics via unified tokens as conditions, it realizes direct human-to-humanoid action transfer. This alignment ensures that human data seamlessly translates into enhanced action controllability for humanoid video generation. Ultimately, by inducing a highly aligned cross-embodiment representation (empirically verified by t-SNE visualizations revealing the convergence of human and humanoid features into a shared manifold), UniT offers a scalable path to distill vast human knowledge into general-purpose humanoid capabilities.

cs.RO↗

arXiv2026

Tools as Continuous Flow for Evolving Agentic Reasoning

Large Language Models (LLMs) have demonstrated remarkable capabilities in orchestrating tools for reasoning tasks. However, existing methods rely on a step-wise paradigm that lacks a global perspective, which causes error accumulation over long horizons and restricts generalization to unseen tools. To overcome these limitations, we propose Tools as Continuous Flow for Evolving Agentic Reasoning (FlowAgent), which reconceptualizes tool chaining as continuous trajectory generation within a semantic space. To systematically evaluate this paradigm, we introduce the first plan-level closed-loop benchmark dedicated to plan-level agentic reasoning in dynamic real-world environments. Specifically, the proposed FlowAgent leverages conditional flow matching to generate continuous latent trajectories, providing a global planning perspective to ensure coherent and robust tool execution. Theoretically, we establish formal bounds on utility convergence and prove that our continuous formulation fundamentally guarantees robust generalization and error attenuation. Empirical evaluations show that FlowAgent achieves superior robustness and adaptability in long-horizon reasoning tasks.

cs.AI↗

arXiv2026

Curv-Tail: Lightweight Long-Tailed Encrypted Traffic Classification with Discrete Packet-Length Encoding and Lorentz Prototypes

Long-tailed encrypted traffic classification requires accurate recognition of infrequent classes under limited computational budgets. We propose Curv-Tail, a lightweight, end-to-end packet--byte framework trained without a separate pretraining stage. Mixed-resolution tokenization preserves exact packet-length identities within a bounded range and coarsens larger values to limit the vocabulary. An auxiliary objective predicts observed length tokens from contextual packet features before pooling, encouraging length-token retention beyond flow-level supervision. Compact temporal encoders process packet sequences and directional byte patches, and Lorentz prototypes with a shared learnable curvature magnitude classify their fused representation. On NUDT-Mobile and DataCon-Website under natural class frequencies, Curv-Tail achieves three-seed mean Tail-F1 scores of 85.70% and 46.30%, exceeding the strongest evaluated baselines by 2.91 and 1.89 percentage points, respectively. In 300-class profiling on an RTX 4090 with FP32 and batch size 256, Curv-Tail uses 98.48% fewer parameters and achieves 10.2 times the batch inference throughput of MM4Flow.

eess.SP↗

arXiv2026

Role-aware Heuristic Episodic Attention for Conversational LLMs

Large language models often lose track of persistent instructions and relevant information as multi-turn conversations grow. We study this cumulative contextual decay through three related failure modes: attention pollution, dilution, and drift. We propose REA (Role-aware Heuristic Episodic Attention), a context-management framework that assigns different persistence and representation policies to instructions and episodic interactions. Instructional Memory retains identified global constraints in a dedicated prefix. Episodic Memory preserves user inputs and compresses model replies, while heuristic retrieval selects raw text, compressed representations, or omission for each historical turn. On Long-MT-Bench+, REA improves the judge score from 6.32 to 7.36 on a 10-point scale, a 16.5% relative gain over the Vanilla baseline, and reduces average latency by 2.91$\times$. Additional evaluations show aggregate gains on three backbones spanning 1.7B-7B parameters and on Chinese and English role-playing tasks. These results support role-aware context management as a practical approach to maintaining conversational continuity and instruction adherence.

cs.CL↗

arXiv2026

FlashBack: Knowing When to Remember in Streaming Vision-Language Models

Streaming vision-language models must process continuously growing video streams under a bounded compute budget, creating a persistent tension between real-time perception and long-term memory. Retrieving historical information provides a natural remedy, yet historical recall is not uniformly beneficial: unnecessary history may introduce irrelevant context into current reasoning and interfere with native real-time perception. Effective streaming memory should therefore address not only what to remember, but also when and how to access it. To this end, we introduce FlashBack, a training-free framework for selective, multi-level memory in streaming vision-language models. Before retrieving history, FlashBack draws on the semantic understanding of the frozen streaming VLM to infer whether a query calls for historical evidence. This assessment determines whether inference remains on the Native trajectory or invokes an isolated Recall trajectory. The Recall trajectory combines recent context with retrieved long-term memory through a query-local Side-KV pathway, preserving local temporal continuity without modifying the persistent Native state. We instantiate FlashBack on StreamingVLM and Mage-VL-4B and evaluate it on OVO-Bench and StreamingBench. The results show improvements on several long-horizon and memory-dependent tasks while largely preserving real-time perception, with performance competitive with strong training-based streaming methods despite requiring no additional training. Our code will be announced later.

cs.CV↗

arXiv2026

First measurement of the one-point charge correlator in $e^+e^-$ collisions at $\sqrt{s} = 91.2$ GeV with DELPHI Open Data

The chiral structure of the $Z$ couplings imprints a parity-odd flow of electric charge on hadronic $Z$ decays. The related forward-backward asymmetries, a key set of observables in the electroweak precision program, were measured at LEP and SLC using the jet charge. The one-point charge correlator offers a complementary route, measuring the hadronic charge flow directly as a function of polar angle relative to the incoming electron-beam axis, without reference to jets or a reconstructed quark direction, following the formalism developed in a companion paper. We report its first measurement, using $61~\mathrm{pb}^{-1}$ of archival DELPHI Open Data recorded at $\sqrt{s} = 91.2$~GeV in 1994 and 1995. Detector effects are corrected in two stages. The first is derived from fully simulated samples, and the second bounds the residual charge-misreconstruction difference between data and simulation using a measurement in $e^+e^-\toτ^+τ^-$ events. The measured charge correlator exhibits the characteristic parity-odd $\sin(2θ)$ modulation and agrees with the \textsc{PYTHIA}~8.3 prediction. The measurement demonstrates that the parity-odd charge flow of hadronic $Z$ decays is directly accessible in data, differentially in polar angle, and the paper discusses strategies for controlling the associated detector effects, paving the way for a new program of charge-flux measurements, both in archival $e^+e^-$ data and at future colliders.

hep-ex↗

arXiv2026

Analysis note: one-point charge correlator with DELPHI Open Data

We present the first measurement of the one-point charge correlator, the angular flux of electric charge in hadronic final states, using DELPHI Open Data collected at LEP-1 at $\sqrt{s} = 91.2$~GeV during 1994 and 1995. The data, corrected for detector effects, exhibit a clear $\sin(2θ)$ modulation, consistent with the parity-violating hadronic charge flow that the chiral structure of the $Z$ couplings imprints on the final state. The measurement demonstrates the experimental feasibility of the observable and establishes strategies for controlling associated detector effects, thereby motivating a new program to measure charge-flux observables. This note documents the experimental details supporting the companion experimental paper and the joint theory--experiment Letter.

hep-ex↗

arXiv2026

Just-In-Time Scene Graph Growth: Combating Perceptual Saturation in Long-Horizon Robotics

While 3D Scene Graphs (3DSGs) provide crucial structured representations for embodied agents, conventional Ahead-of-Time, "build-everything-then-filter" pipelines conflict with the real-time, low-latency demands of edge platforms, inducing a perceptual saturation effect via severe observation redundancy. To resolve this, we present JITOMA (Just-In-Time On-demand Memory Activation), a closed-loop framework that unifies task reasoning, perception, and memory into a just-in-time growth process. Instead of exhaustively mapping the entire environment, JITOMA leverages a top-down task heatmap at the frontend to filter continuous observations, routing minimal streams to maintain a global foundation of low-cost, dormant anchors. Upon a cognitive query, the backend Large Language Model (LLM) parses the robotic intent to dynamically awaken task-relevant anchors, triggering expensive semantic operations such as dense node captioning exclusively within the activated local subgraph. To evaluate these dynamic capabilities and study perceptual saturation trade offs, we introduce JITOMA-Bench, a benchmark for long-horizon task switching and complex intent grounding. Across JITOMA-Bench, JITOMA maintains only 1--6 active semantic objects and 0.25--0.28 s/frame across all tiers, showing that semantic computation remains bounded by current task demand rather than accumulated scene complexity.

cs.CV↗

arXiv2026

Resolving State-Representation Mismatch: State-Space Visual Reasoning for Open-Loop VLA Planning

Despite rapid progress in vision-language-action (VLA) models, existing reasoning paradigms still face a fundamental \emph{state-representation mismatch} in open-loop planning. Given only an initial observation, models must internally simulate action-conditioned state transitions, whereas text-, pixel-, and latent-space reasoning can suffer from lossy spatial compression, error-accumulating visual generation, and bypass of intermediate latent tokens, respectively, undermining reliable long-horizon planning. We propose \textbf{State-Space Visual Reasoning} (SSVR), which decouples static visual context, language constraints, and a recurrent latent state. SSVR encodes the initial image and instruction once, then conditions each action prediction on the latent state and updates it with an action-conditioned GRU. Using Qwen2.5-VL as the backbone, SSVR achieves 99.5/99.6, 96.3/98.0, and 83.9/90.6 EM/PR on FrozenLake, Maze, and MiniBehavior, substantially outperforming prior methods. Extensive experiments support the effectiveness of recurrent state modeling for VLA open-loop planning across input transformations and transfer settings. By reusing static visual-textual context and updating a compact recurrent state, SSVR supports efficient multi-step inference, achieving up to $98.58\times$ faster Maze decoding rollouts than the evaluated baselines with the prefix cache prebuilt.

cs.CV↗

arXiv2026

OpenFC: Learning Verification Policies towards Open-Search Fact Checking

Open-search fact checking is not merely retrieval followed by classification, but a sequential decision problem in which every query, source visit, and stopping decision reshapes the evidence available for verification. Yet existing systems often distribute these decisions across predefined pipelines or separately prompted modules rather than learning them as a unified task-specific policy. We introduce \textbf{OpenFC}, a unified verification-policy training framework that post-trains Qwen3-8B as a compact next-action controller over reasoning, evidence acquisition, and stopping. OpenFC learns this policy in two stages. \textbf{Stepwise-Calibrated Cold Start (SCCS)} uses a strong training-time supervisor to review post-initial reasoning, tool-use, and stopping proposals before execution, producing reliable trajectories for supervised fine-tuning without access to gold verdicts. \textbf{Verification-Aware Reinforcement Learning (VA-RL)} then improves the cold-start policy on unresolved claims through budget-aware tool rewards, label-aware advantage reweighting, and localized response masking. Across six fact-checking benchmarks, OpenFC achieves 70.39\% average accuracy and 63.30\% macro-F1, the highest overall averages among the evaluated methods. Stage-wise ablations further show that SCCS and VA-RL provide complementary gains, supporting the design of the two-stage training framework. These results position OpenFC as a strong and effective framework for open-search fact-checking. We will open-source our code and release the model checkpoints to support reproducibility.

cs.AI↗

arXiv2026

Dual-Frontier: When Can an Agent Trust Its World Model?

Learned world models are becoming essential to general-purpose agents: by predicting action consequences, they support planning and decision-making while reducing reliance on costly trial and error. This reliance creates a fundamental ambiguity: when a world-model-guided decision fails, the trajectory alone may not reveal whether the agent's decision rule or the world model caused the loss. We formalize this failure-attribution problem as a counterfactual decomposition of return loss and prove that its components are not identifiable from passive interaction, even for finite-horizon planners. This obstruction motivates Dual-Frontier, a learning principle that admits a world-model-guided decision only when its predicted advantage exceeds a certified bound on decision-relevant world-model error; otherwise, evidence is allocated to world-model verification. Action-conditioned value bounds and a closed-loop extension guarantee non-decreasing return for admitted decisions. Calibrated gates and simultaneous confidence sequences support adaptive evidence reuse, with sufficient and necessary verification bounds. Controlled learned-model experiments validate the predicted failure modes and certification behavior, while cross-backbone tool-use benchmarks instantiate the same verify-then-promote rule in realistic agent world-model pipelines, consistently improving decision quality and reliability.

cs.AI↗