arXiv subjects

Xi Chen

Publications and source records attributed to Xi Chen.

At least 55 records · Page 3Linked to original sources

arXiv2026

GRAM: Correcting Frozen Time-Series Foundation Models via Graph-Retrieved Amplitude Memory

Time-series foundation models (TSFMs) enable zero-shot forecasting through large-scale cross-domain pretraining, while retrieval augmentation further improves their performance by leveraging historical information. However, existing methods typically correct TSFM forecasts using the ground-truth futures of similar historical windows, which contain both predictive components already captured by the foundation model and sample-specific random fluctuation that is difficult to transfer. In contrast, recurring systematic model bias within prediction errors more directly characterizes the failure modes of a frozen TSFM and therefore provides more valuable correction signals. Effectively exploiting such model bias, however, poses two challenges: prediction errors at different numerical levels are difficult to compare due to scale differences, and the recurring bias must be extracted from prediction errors contaminated by random fluctuation. To address these challenges, we propose GRAM, a general retrieval-augmented framework for frozen TSFMs. GRAM first introduces an Amplitude Memory Module (AMM) that scales prediction errors by amplitude and aggregates them into retrievable prototypes. It then employs a Prototype Graph Module (PGM) to model relations among prototypes to aggregate consistent bias information while suppressing random fluctuation. During online forecasting, GRAM retrieves and expands prototypes relevant to the current query and generates per-horizon corrections to refine the original TSFM forecast. Experiments across multiple datasets and foundation models demonstrate consistent forecasting improvements.

cs.LG↗

arXiv2026

Throttling Equilibria in Auction Markets

Throttling is a popular method of budget management for online ad auctions in which the platform modulates the participation probability of an advertiser in order to smoothly spend her budget across many auctions. In this work, we investigate the setting in which all of the advertisers simultaneously employ throttling to manage their budgets, and we do so for both first-price and second-price auctions. We analyze the structural and computational properties of the resulting equilibria. For first-price auctions, we show that a unique equilibrium always exists, is well-behaved and can be computed efficiently via tatonnement-style decentralized dynamics. In contrast, for second-price auctions, we prove that even though an equilibrium always exists, the problem of finding even an approximate equilibrium is PPAD-complete, there can be multiple equilibria, and it is NP-hard to find the revenue maximizing one. We also compare the equilibrium outcomes of throttling to those of multiplicative pacing, which is the other most popular and well-studied method of budget management. Finally, we characterize the Price of Anarchy of these equilibria for liquid welfare by showing that it is at most 2 for both first-price and second-price auctions, and demonstrating that our bound is tight.

cs.GT↗

arXiv2026

Text-Video Retrieval via Multi-Dimensional Saliency Assessment and Granularity-Aware Query Decomposition

Text-video retrieval, which aims to bridge visual and textual modalities by learning a joint embedding space, has become a crucial task in multimodal intelligence. Despite extensive efforts to mitigate visual redundancy, previous methods typically rely on a single-aspect criterion to assess visual importance, overlooking the multifaceted spatiotemporal nature of video. In addition, encoding text into a single global embedding to align with videos compresses temporal events and spatial entities into a unified representation space, further aggravating cross-modal misalignment. To address these issues, we propose MMTI, a method that jointly mitigates visual redundancy and enables multi-grained text-video interaction to achieve accurate multi-grained semantic alignment. Specifically, a key feature selection (KFS) mechanism adaptively identifies and aggregates informative frames and patches by jointly evaluating multi-dimensional saliency and learnable importance scores, effectively compacting dense visual features and mitigating visual redundancy. Furthermore, our proposed multi-grained text-video interaction module (TVIM) employs a dynamic gating mechanism to decompose the text query into sentence, frame, and patch queries (SFP), enabling multi-grained text-video alignment. Complementary alignment at different granularities is thereby achieved. Extensive experiments on four standard benchmarks demonstrate that our method outperforms state-of-the-art methods.

cs.IR↗

arXiv2026

Two-Fidelity Best-Action Identification for Stochastic Minimax Tree

We study fixed-confidence best-action identification (BAI) in stochastic minimax trees. This problem is increasingly relevant in modern AI planning, where deep minimax search and Monte Carlo Tree Search (MCTS) with language model long rollouts face a fundamental tradeoff: heuristic evaluations are cheap but biased, while accurate rollouts are reliable but prohibitively expensive. We propose 2FFS, a two-fidelity tree-search algorithm that brings multi-fidelity flat bandit ideas into trees. The algorithm combines minimax-style fast expansion with MCTS-style stochastic sampling, adaptively deciding when to exploit cheap biased evaluations and when to invoke expensive accurate evaluations for local certification. We prove fixed-confidence correctness, establish finite stopping for exact identification, and give a polynomial-depth cost upper bound for general-depth trees. Across numerical stochastic-tree experiments, 2FFS uses substantially fewer samples and computational operations comparing to existing BAI-MCTS baseline.

cs.LG↗

arXiv2026

AnthroDial: Benchmarking LLM Anthropomorphism in Autonomous Social Interaction

Large language models (LLMs) are increasingly deployed as social agents, yet credible human-like interaction requires more than fluent responses or persona consistency. Agents must autonomously decide whether, when, and how to communicate while adapting to evolving contexts, goals, and relationships. Existing research, however, lacks a unified approach to enabling, evaluating, and improving such capabilities in continuous, open-ended interaction. We introduce AnthroDial, a unified framework for developing anthropomorphic social agents from three complementary aspects: MindFlow, a lightweight interaction harness that enables autonomous, asynchronous, and adaptive communication through a dynamic Mind Buffer; CAPS-Eval, a theory-grounded framework for evaluating cognitive, affective, and behavioral dimensions of anthropomorphic interaction; and a scalable training paradigm that combines SEEDS for environment expansion with DiAPO for adaptive capability optimization. We further construct evaluation datasets covering everyday communication, game interaction, and long-horizon character interaction. Extensive experiments across diverse models and scenarios demonstrate improved interaction autonomy and naturalness, validate the reliability, discriminativeness, and agreement with human rankings of CAPS-Eval, and confirm the effectiveness of our training paradigm. Together, these components provide a unified framework for developing credible human-like social agents in open-ended interaction.

cs.CL↗

arXiv2026

Relevance-Resolution Transfer via Scale-Decomposable Fractional Diffusion for Multi-Length Cross-Modal Hash Retrieval

Cross-modal hashing enables efficient retrieval by encoding heterogeneous data into compact binary codes. Recent methods exploit fine-grained relations encoded in multi-label training structure, yet none of them constrains how those relations survive as consistent candidate rankings in finite, multi-length Hamming spaces, which we term the relevance resolution bottleneck (RRB). To address the RRB, we propose MultiBit, which transfers relevance resolution from multi-label structure to multi-length Hamming spaces. MultiBit first constructs a scale-decomposable fractional relation teacher from dataset-level label co-occurrence and label specificity, and models dependencies from local to long-range over continuous diffusion scales. It then maps the discretized diffusion scales and their quadrature weights to scale-aware bit subblocks of the maximum-length code, organizes the target code lengths as nested prefixes, and aligns their Hamming candidate rankings with the teacher relations. Experiments on multiple benchmarks demonstrate improved retrieval accuracy. Code is available in the supplementary material.

cs.IR↗

arXiv2026

PanoVLN: Towards Effective Panoramic Vision-and-Language Navigation

Recent vision-language models (VLMs) have advanced vision-and-language navigation (VLN), enabling models to predict navigation actions from visual observations and language instructions. In this work, we explore VLN with panoramic observations and introduce PanoVLN. The motivation is straightforward: more complete visual context should enable better-informed navigation decisions. For example, a panorama can reveal a passage outside a perspective camera's field of view, allowing the model to identify the intended route without additional exploration. However, we find that simply replacing perspective images with panoramas yields only limited gains. Our diagnosis suggests that fully exploiting wider visibility requires modifications to action prediction, training supervision, and visual representation. First, wider visibility supports longer-horizon action planning. We make the model predict longer action sequences, enabling larger turns and subsequent movement from a single panorama. Specifically, we introduce a confidence-guided execution (CGE) strategy that dynamically determines how many predicted actions to execute before replanning. Second, wider visibility also brings more complex route choices. We therefore construct training routes with frequent branching points and clear instructions to provide targeted supervision for route selection. Third, panoramic navigation requires understanding spatial relationships across viewing directions, beyond recognizing individual landmarks. We combine semantic and geometric features from RGB panoramas to capture both scene content and spatial layout without adding visual tokens. With a 4B backbone and RGB-only input, PanoVLN surpasses the previous SOTA by 11.9% and 8.7% in success rate on R2R-CE and RxR-CE Val-Unseen. Real-world experiments on a quadruped further demonstrate faster navigation with fewer pauses than prior VLN methods.

cs.CV↗

arXiv2026

When Words Speak Louder than Images: Towards Understanding Language Bias in Vision-Language Models

Despite substantial progress across downstream applications, vision-language models (VLMs) remain susceptible to language bias, often prioritizing linguistic cues over visual evidence and consequently producing incorrect predictions. Prior studies have proposed various approaches to understanding and mitigating language bias in VLMs, yet their findings often conflict due to the difficulty of tracing how language bias propagates within black-box VLMs. Building on the word completion task, we trace how language bias propagates through VLM inference by (1) proposing a diagnostic framework that decomposes the inference process into four distinct yet interdependent stages to trace the propagation of language bias; and (2) examining how two key factors underlying language bias, i.e., linguistic priors and cross-modal coverage, evolve across these stages and ultimately give rise to incorrect predictions. The linguistic prior captures the strength of statistical bias induced by the language model component of a VLM and represents the origin of language bias, whereas cross-modal coverage measures the extent to which linguistic cues cover the visual content. By decomposing inference into four stages and characterizing the interplay between linguistic priors and cross-modal coverage across these stages, we propose a systematic framework for tracing the propagation of language bias throughout the inference process; and uncover the underlying mechanism of language bias by revealing the interplay between linguistic priors and cross-modal coverage.

cs.CL↗

arXiv2026

MLLMs Know When Before Speaking: Revealing and Recovering Temporal Grounding via Attention Cues

Video temporal grounding (VTG), which localizes the start and end times of a queried event in an untrimmed video, is a key test of whether multimodal large language models (MLLMs) understand not only what happens but also when it happens. Although modern MLLMs describe video content fluently, their timestamp predictions remain unreliable, while existing remedies either require costly post-training on temporal annotations or rely on coarse training-free heuristics. In this work, we probe the cross-modal attention of MLLMs and uncover a perception-generation gap. Our key finding is that MLLMs often know the target interval during prefill, but lose this signal when generating the final answer. In the prefill stage, a sparse set of attention heads, which we call Temporal Grounding Heads (TG-Heads), concentrates query-to-video attention on the ground-truth interval. During autoregressive decoding, however, the answer tokens shift attention away from this interval toward visually salient but query-irrelevant segments. This observation motivates an inference-time read-then-regenerate framework. We first convert TG-Head prefill attention into a debiased frame-level relevance signal and extract the high-attention interval it highlights. We then re-invoke the MLLM with visual context restricted to this interval using video cropping. Without parameter updates or architectural changes, our framework consistently improves MiMo-VL, Qwen3-VL, and Molmo2 on three benchmarks, with gains of up to +3.5 mIoU. The project website can be found at https://ddz16.github.io/mllmsknowwhen.github.io/.

cs.CV↗

arXiv2026

Learning Spatiotemporal Sensitivity in Video LLMs via Counterfactual Reinforcement Learning

Video large language models (Video LLMs) can achieve strong video-QA accuracy without reliably tracking spatiotemporal dynamics. A model may answer a motion question from static cues, for example, and give the same prediction even after the underlying motion is reversed. Correctness-based reinforcement learning does not directly address this problem because it rewards the final answer without requiring the policy to respond to task-relevant video dynamics. We propose Counterfactual Relational Policy Optimization (CRPO), which explicitly trains Video LLMs to respond to controlled changes in the visual input. For each training example, CRPO constructs a counterfactual video, such as a horizontally flipped or temporally reversed version of the original, and jointly optimizes rollouts from both videos under a shared policy. Factual supervision anchors what the model should answer, while counterfactual supervision constrains when that answer should change: predictions should change when an intervention alters task-relevant dynamics and remain stable when the queried property is preserved. This coupling provides a direct behavioral learning signal without requiring ground-truth labels for transformed videos or annotated spatiotemporal reasoning traces, while discouraging indiscriminate answer changes. To evaluate this property, we introduce DyBench, a paired counterfactual benchmark with 3{,}014 videos and a strict pair-accuracy metric. Across paired spatiotemporal evaluations and standard video benchmarks, CRPO improves sensitivity to motion and temporal changes while improving performance on general video understanding. The gains also extend to segment reordering, a transformation never used during training, suggesting that CRPO learns sensitivity to video dynamics beyond the training interventions. The project website can be found at https://ddz16.github.io/crpo.github.io/ .

cs.CV↗

arXiv2026

Can a Robot Read Braille? - Learning to Adapt Contact via Imitation Learning for Tactile Braille Recognition

For people who are blind, touch provides an essen-tial channel for accessing written information through Braille. Bringing a similar capability to robots requires them not only to recognize tactile patterns, but also to actively establish physical contact that makes those patterns readable. Yet existing robotic Braille readers largely focus on recognition after contact, leaving contact establishment itself insufficiently addressed. We present an adaptive-contact framework for robotic tactile Braille reading that assesses contact quality and physically corrects unsuitable contact before recognition and reconstruc-tion. Multi-Head Policy Learning uses expert-guided contact-adjustment demonstrations to jointly learn contact acceptability and pose corrections. During deployment, the robot iteratively evaluates and re-establishes contact, retaining reliable tactile observations for pose-aware fusion and Braille reconstruction. Across 20 physical Braille plates used for learning and eval-uation, the proposed approach achieves 94.0% tactile quality and 88.6% tactile reconstruction on the ten online-evaluation plates. These results demonstrate the importance of actively establishing readable contact, rather than relying solely on recognition under imperfect tactile observations, for reliable robotic Braille reading.

cs.RO↗

arXiv2026

VLALight: Lightweight Vision-Language-Action Models for Emergency-Aware Traffic Signal Control

Traffic signal control (TSC) is essential for mitigating urban congestion. Recent advances in vision-language models (VLMs) enable richer interpretation of intersection scenes, opening new opportunities for visual-context-aware TSC. However, the loose coupling and repeated information conversion between modules can lead to the loss of fine-grained visual details, while sequential inference introduces substantial latency. To address these limitations, we propose VLALight, a lightweight end-to-end vision-language-action framework that directly maps intersection observations and signal-phase information to discrete signal actions. To handle the multi-view nature of TSC, VLALight combines multiple directional camera views into a unified visual input and uses textual instructions to establish their correspondence with traffic movements and signal phases. This design enables direct action prediction with a compact 0.5 B-parameter model, without intermediate image-to-text descriptions or handcrafted traffic-state representations. Experiments show that VLALight delivers the best emergency-vehicle service of all compared methods, reducing pooled emergency waiting time by 21.1% over the cascaded VLMLight while running in real time on local hardware and generalizing to unseen intersection topologies and traffic-flow patterns.

cs.CV↗

arXiv2026

Attention Sinks and Outliers in Attention Residuals

We propose OASIS, an outlier- and sink-aware method that stabilizes dual-normalized attention-residual architectures through explicit null routing and token-to-depth null coupling. AttnResidual introduces an additional depth-wise normalization channel that improves inter-layer routing flexibility but can also amplify attention sinks, activation outliers, and low-bit quantization error. OASIS builds on explicit Softmax1-based null routes at both the token and depth levels and uses token-level null evidence to downweight depth branches exhibiting stronger null behavior. Theoretically, we characterize a conditional mechanism for sink-like attention concentration under dual normalization, offering insight into the low-bit sensitivity observed in AttnResidual. Experimentally, we compare OASIS against five baselines on three language-model backbones and multiple language-modeling, reasoning, and long-context benchmarks and observe consistent improvements in both attention sink mitigation and post-quantization performance. Notably, relative to Vanilla AttnResidual, OASIS reduces maximum infinity norm by 81.75% and average kurtosis by 95.90%, lowers W8A8 perplexity by 82.00%, and improves W4A4 GSM8K Pass@1 by 42.11% on average across LLaMA-3.2-1B, Qwen3-0.6B, and Phi-4. Code is available at: https://github.com/robinzixuan/OASIS.

cs.LG↗

arXiv2026

Antichain polynomials of products of chains and minuscule posets

This paper studies the antichain polynomials of $[k]\times P$, where $P$ is a connected minuscule poset. We give a formula for the number of antichains, counted by size, of an arbitrary poset. Using this formula, we present necessary and sufficient conditions for the palindromicity of antichain polynomials for two infinite families of connected minuscule posets. We show that, for every connected minuscule poset $P$, if the antichain polynomial of $[k]\times P$ is palindromic, then it has only real and strictly negative zeros. This result, in particular, gives an affirmative answer to Ding-Dong's conjecture about $γ$-positivity of the antichain polynomial of $[k]\times P$. By constructing a bijection between antichains and labeled Dyck paths, we also give a layer-refined enumeration of the antichains in $[2]\times[m]\times[n]$. We establish a relation between such antichains and Clar covers of the hexagonal flakes $O(2,m,n)$, thus prove a conjectured determinantal formula for a family of Zhang-Zhang polynomials. Real-rootedness and stability results are also obtained when the shortest chain has length at most two. Finally, we present infinitely many connected Peck posets whose antichain polynomials are not unimodal, disproving the log-concavity conjecture of Ding and Dong.

math.CO↗

arXiv2026

Antisymmetric breathing in altermagnetic skyrmions

A skyrmion in an altermagnet with $d$ wave symmetry consists of two elliptical sublattice textures with perpendicular long axes. For each sublattice component $η=A,B$, we define an effective skyrmion radius $R_η=\sqrt{a_ηb_η}$, where $a_η$ and $b_η$ are the distances from the skyrmion center to its boundary along $y$ and $x$, respectively. For the anisotropic exchange parameters studied, the equilibrium radius at zero field is smaller than in a reference with parallel sublattice textures and otherwise identical parameters. A magnetic field perpendicular to the film expands one sublattice skyrmion and contracts the other, generating a radius difference $\dR=R_A-R_B$ and a net magnetic moment. The exchange modulation used to represent uniaxial strain also produces a nonzero radius difference at zero field. Because the strong and weak exchange directions are interchanged between the two sublattices, a common directional change of the exchange couplings increases one radius and decreases the other. Unlike the magnetic coupling, this mechanism vanishes when the two sublattice exchange tensors become identical. The radius difference also supports an antisymmetric breathing mode. After a short field pulse, it oscillates in quadrature with the uniform helicity, the common rotation of the wall magnetization within the film plane, at $49.5\,\mathrm{GHz}$ for the reference parameters.

cond-mat.mes-hall↗

arXiv2026

Combinatorial aspects of the Delannoy Lattice

The Delannoy numbers $d(n,k)$ count lattice paths from $(0,0)$ to $(n-k,k)$ using steps $(1,0),(0,1)$ and $(1,1)$. This paper introduces a graded poset $\mathcal{D}_n$ on the Delannoy paths ending on the line $x+y=n$, whose rank-generating function is the Delannoy polynomial $d_n(x)=\sum_{k=0}^n d(n,k)x^k$. We prove that $\\mathcal{D}_n$ is a self-dual lattice, which is call the Delannoy lattice. By establishing an explicit symmetric Boolean decomposition, we show that $\mathcal{D}_n$ is a symmetric Boolean order, thereby recovering the $γ$-positivity of $d_n(x)$. Such a decomposition is refined to a symmetric chain decomposition with the chain cover property, and is applied to determine all maximum antichains. We also investigate other combinatorial aspects of $\mathcal{D}_n$, including supersolvability, the Möbius number, characteristic polynomials, and zeta polynomials.

math.CO↗

arXiv2026

Logarithmic concentration for the focusing cubic Schrödinger equation on the two-dimensional torus

We construct a smooth solution of the focusing cubic nonlinear Schrödinger equation on a square two-dimensional torus which is global in both time directions and concentrates at one point as time tends to positive infinity. Its mass is the mass of the Euclidean ground state and its energy is zero. If $λ(t)$ denotes its concentration scale and $L$ the period, then $L/λ(t)=2\log t+\frac{11}{2}\log\log t+O(1)$. The solution approaches a phased, periodized ground state with an error tending to zero in every fixed Sobolev space. In particular, its $H^s$ norm is asymptotic to an explicit constant times $(\log t)^s$ for every $s>0$. We also prove uniqueness, up to constant phase and time translation, within the quantitative single-profile asymptotic class used in the construction. The mechanism is the exponentially small interaction of the profile with its periodic images. Exact mass and energy constraints select the concentrating branch, while a positive energy on the moving constrained space controls the terminal-value problem at every fixed regularity order.

math.AP↗

arXiv2026

Co-Fabric: Breaking Host-Domain Boundaries for Unified xPU Interconnection

Large-model parameters have grown beyond the capacity of a single xPU, dispersing across multiple xPUs spanning distinct host domains, where xPU-to-xPU communication dominates overall system efficiency. Existing scale-up interconnect remains inadequate: network-based solutions built on Ethernet--such as RoCE (RDMA over Converged Ethernet)--introduce specific message-semantics and protocol-stack characteristics, and rely on fragmented per-host addressing, while conventional host-based fabrics are confined to a single host domain and lack cross-host unified addressing. This paper presents Co-Fabric, a bus-based interconnect that, unlike conventional bus designs, breaks host-domain boundaries to deliver unified xPU interconnection for scale-up superpods. Co-Fabric makes three contributions: a streamlined four-layer protocol stack achieving nanosecond-scale processing latency with native reliability; a cross-domain scaling and P2P mechanism that routes using port identifiers embedded in the packet header; and a unified address space built on shadow-device auto-enumeration. On a 64-xPU 3D-Mesh system, Co-Fabric cuts inter-node communication latency by over 50% and improves bandwidth by 2-5x over RoCE, accelerating DeepSeek R1 inference by 30%-80%. Moreover, since its streamlined four-layer protocol stack and higher data-communication efficiency reduce protocol and processing overhead relative to the Ethernet-based RoCE stack, Co-Fabric cuts the cost and power of the interconnect itself by up to 80% and 5%, respectively. These results demonstrate Co-Fabric's advantage for AI computing centers.

cs.DC↗