跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+197新增论文
+151其中 AI 相关
+16已有论文修订
+91新增 AI 导读

arXiv 最近完成扫描:2026-11-21 12:19 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-21 03:55 — 2026-11-22 03:55(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-21 12:19最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 21,490 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-21 12:17(北京时间)
新论文 · v1arXiv:2611.21921

Latent Fact Composition under Next-Token Training

Tianyi Zhang, Feiying Chen

原文摘要

Controlled studies of latent two-hop reasoning train transformers to predict only the answer, whereas language models are trained to predict every token. We ask whether a model trained on every token learns to answer two-hop questions whose intermediate entity appears in no two-hop training example. We build a synthetic corpus with such questions. On this corpus every term of the loss can be computed exactly. A looped transformer, a standard transformer and DiscoLoop answer almost none of these questions. We therefore introduce the Baton family of looped transformers. They pass the intermediate entity to the second loop as a token embedding. When trained to predict only the answer, two of its three variants answer these questions almost perfectly. We then relate the success and failure of each model to three design conditions. Experiments that change one condition at a time agree with this hypothesis. When Baton is trained on every token, however, its accuracy on these questions rises and then declines. We further split the gradient of the next-token loss into three terms. Removing the two terms that depend on the particular training set raises the peak accuracy but does not remove the decline. The decline arises in the second hop while the model still recalls both facts.

cs.CL · cs.LG

版本更新 · v2arXiv:2608.27754

Towards Efficient Auto-Interpretability of Boltz-1 Sparse Autoencoders

Piotr Jedryszek, Oliver M. Crook

原文摘要

Sparse autoencoders (SAEs) could turn AI models in Biology and other fields into engines of scientific discovery by explaining the superhuman capabilities of those models. We postulate that a latent is useful if we know three things: whether it is coherent, whether it can be described, and whether that description has predictive power. These questions are routinely conflated. We assemble them into a single pipeline and report some practical innovations. First, cross-seed dictionary stability prioritises which latents are worth spending resources to investigate. Second, an intruder-detection task asks whether a latent's activating examples share a recognizable pattern. Third, a separate pass proposes a candidate biological description which we convert into falsifiable predictions which can be tested in silico. Deployed on the Boltz-1 Pairformer trunk, stability prioritisation finds interpretable latents using about 4.4 times fewer latent evaluations each, and at 5.2 times lower measured LLM cost, while recovering over half of them. However, we notice that in our setup the LLM judge is heavily biased towards consistent residue identity and a simple filter that selects for features with high amino-acid identity precision outperforms the stability thresholding. This simpler method has the added benefit of requiring no extra SAE seeds but it remains uncertain if it reflects a genuinely useful heuristic or an limitation of the LLM judge.

q-bio.QM · cs.AI

新论文 · v1arXiv:2611.21920

KVUse: Reusable-KV User Memory for Efficient LLM Recommenders

Xin Guo, Qiang Zhang, Xiang Zhou, Arushi Rai, Siyu Lin, Yifan Wu, Xiang Shen, Lei Huang 等 11 位作者

原文摘要

Memory-augmented recommenders built on large language models (LLMs) maintain a natural-language profile of each user. These profiles are updated by a user profile update path that revises the profile as new interactions arrive and are consumed by downstream task paths like ranking. Because these user profiles are stored as text in the user memory, every ranking LLM call that consumes them must splice them in the prompt and calculate the KV cache of them during the prefill phase. Given these ranking LLM calls are prefill heavy (long prompt but short decoded output) and repeatedly called, these repeated prefill of user profile text dominates the serving. Prefix caching cannot remove it, because these user profiles are preceded by different text in different calls. We present KVUse, which stores the user profiles as reusable key-value (KV) caches instead of text. KVUse trains lightweight LoRA writers by self-distillation, so that the user profile update call produces the reusable KV of the user profile while decoding it, and the LLM can read these reusable KV at new positions and behind new context without recomputation. KVUse also has a customized KV scheduler that uploads the stored reusable KV of the next batch while the current batch is computed, hiding the KV transfer from host memory behind GPU computation. On four InstructRec datasets with Llama-3.1-8B-Instruct, KVUse shows no statistically significant difference in ranking quality from full prefill over continuous memory update rounds. Under saturated load, KVUse raises the throughput of the downstream scoring path by 4.0x-4.5x and of the profile update path by 1.4x-1.9x.

cs.LG

新论文 · v1arXiv:2611.21919

ALOHA Lightning: Learning Fast and Precise Manipulation

John Hua Yao. Qi Wu, Yihuai Gao, Chelsea Finn, Zipeng Fu

原文摘要

Learning from human demonstrations has enabled robots to acquire a wide range of manipulation skills, but learned policies typically execute far slower than ordinary humans. This speed gap is mainly due to lack of an interface for collecting demonstration data at high speed, and the difficulty in training policies that can robustly execute high-speed motions. In this paper, we present ALOHA Lightning (Flash ALOHA), a system for learning fast and precise robotic manipulation. Our system uses kinesthetic teaching to intuitively collect near-human-speed demonstrations on a backdrivable bimanual platform, yielding natural and fast trajectories. We also present a learning pipeline that enables smooth high-speed execution through test-time action smoothing and aligns the visual data distribution between data collection and deployment with masking. Given 50 demonstrations for each task, ALOHA Lightning autonomously completes bimanual tasks such as ball tossing and catching, folding shorts, battery insertion, and bussing tables for over 80% success rates while close to human speed. Project website: https://flash-aloha.github.io/

cs.RO · cs.AI · cs.LG · eess.SY

新论文 · v1arXiv:2611.21917

Lost in a Few Heads: Diagnosing and Repairing Block Selection in Sparse Attention

Wenhe Zhang, Boxing Chen

原文摘要

Sparse attention models read a few blocks of keys for every query at a fixed budget and can lose much of their long-context retrieval accuracy at that budget. We ask where this loss is made. Using exact attention mass selection at the shipped budget as an oracle, we show that the selector under-ranks the needle rather than running out of budget. We find attention heads that miss the needle by running a calibration probe that teacher-forces the answer and reads two quantities in every head. The first is the needle attention that the selector forfeits, and the second is the share of the answer's log-probability that fixing the head's selection restores, integrated along the path from the shipped selector to full attention. Four checkpoints of the InfLLM-V2 family and NOSA-8B lose multi-key retrieval in the same three blind heads, which attend to the needle while a mean-pooled key summary dilutes its single decisive token. Full attention in those heads or a 4-bit scorer that keeps the shipped budget recovers most of the loss without training. On MiniCPM4.1-8B at 64K, either fix lifts multi-key retrieval accuracy from 54.1 to 97.5%, level with full attention in every head. In production serving engines, full attention in those heads keeps 84 to 89% of the time to first token (TTFT) that sparse attention saves at 32K to 128K context length. Beyond the InfLLM-V2 selector lineage, one unified rule flags attention heads above a threshold swept on each checkpoint until they close most of the gap to full attention on calibration prompts. A similar small set of blind attention heads carry the loss for the SeerAttention-R gates, NSA, and DashAttention, whereas many such heads exist in MoBA. On Qwen3-8B with the SeerAttention-R gates, full attention in 18 of 288 key-value heads lifts multi-key retrieval accuracy from 29.8 to 97.7%.

cs.LG

版本更新 · v2arXiv:2604.05102

Finite-Step Invariant Sets for Hybrid Systems with Probabilistic Guarantees

Elizabeth Dietrich, Varun Madabushi, Murat Arcak, Hanna Krasowski, Maegan Tucker

原文摘要

Poincaré return maps are a fundamental tool for analyzing periodic orbits in hybrid dynamical systems by reducing the stability analysis of a periodic orbit to that of a discrete-time system. While linearization provides local stability information, assessing robustness to disturbances requires identifying invariant sets of the state space under the return dynamics. However, computing such invariant sets is computationally difficult, especially when system dynamics are only available through forward simulation. In this work, we propose an algorithmic framework leveraging sampling-based optimization to compute a finite-step invariant level set around a nominal periodic orbit using sampled evaluations of the return map. The resulting solution is accompanied by probabilistic guarantees on finite-step invariance. We provide theoretic analysis establishing sufficient conditions for which the algorithm terminates and demonstrate the approach on two low-dimensional systems and a Compass-Gait Walking model.

eess.SY · cs.RO

版本更新 · v3arXiv:2604.21657

Transferable SCF-Acceleration through Solver-Aligned Initialization Learning

Eike S. Eberhard, Viktor Kotsev, Timm Güthle, Stephan Günnemann

原文摘要

The cost of Kohn-Sham density functional theory (KS-DFT) calculations scales with the number of required solver iterations, which depend on the quality of the initial guess. Machine learning methods that predict initial guesses from molecular geometry can reduce this cost, but matrix-prediction models fail when extrapolating to larger molecules, degrading rather than accelerating convergence [Liu et al., 2025]. We show that this failure is a supervision problem, not an extrapolation problem: models trained on ground-state targets fit those targets well out of distribution, yet produce initial guesses that slow convergence. Solver-Aligned Initialization Learning (SAIL) resolves this for both Hamiltonian and density matrix models by differentiating through parts of the self-consistent field (SCF) solver end-to-end. We introduce the Effective Relative Iteration Count (ERIC), a correction to the commonly used RIC that accounts for hidden Fock-build overhead. On QM40, which contains molecules up to 4$\times$ larger than the training distribution, SAIL reduces ERIC by 40% (PBE), 35% (SCAN), and 32% (B3LYP), more than doubling the previous state-of-the-art reduction on B3LYP. On QMugs molecules 10$\times$ larger than the training set, SAIL delivers a 1.35$\times$ wall-time speedup at the hybrid level of theory, extending ML SCF acceleration to large drug-like molecules.

cs.LG

新论文 · v1arXiv:2611.21907

REEL: Brain-Guided Video Generation and Selection by Reranking Encoder Embeddings with Linear Decoding from fMRI data

Fabrizio Spera, Tommaso Boccato, Moein Khajehnejad, Nicola Toschi, Matteo Ferrante

原文摘要

Visual brain decoding from functional magnetic resonance imaging (fMRI) data has made remarkable progress in recent years, driven by increasingly powerful artificial intelligence (AI) models and the growing availability of large-scale public datasets. However, despite the impressive advances achieved in static image decoding, the reconstruction of videos from fMRI signals remains challenging due to the temporal mismatch between rapidly changing video stimuli and the slow dynamics of the blood-oxygen-level-dependent (BOLD) response. We evaluate a Bayesian-inspired proposal-and-rerank strategy on CC2017 and the BOLD Moments Dataset (BMD), using a linear semantic decoder, a pretrained video generator, and three neural encoding streams with ridge readouts and learned pooling in two streams. We refer to this framework as REEL (Reranking Encoder Embeddings with Linear decoding). Compared with a single generated candidate, reranking increases mean Video 50-way accuracy from 0.2359 to 0.2434 across three CC2017 participants and from 0.3653 to 0.3920 across ten BMD participants. These findings suggest an alternative and conceptually simpler strategy for improving video brain decoding, highlighting the potential of our approach as a complementary direction to increasingly complex pipelines.

cs.CV

版本更新 · v3arXiv:2602.11476

Bounded Local Generator Classes for Deterministic State Evolution

R. Jay Martin

原文摘要

Persistent memory must retain history without revisiting it at every transition. Bounded Local Generator Classes (BLGC) formalize carried numerical state, selected context and ordered local updates. The $\ell^\infty$ formulation permits countably many uniformly bounded coordinates without square summability. Projected maps preserve admissibility, Lipschitz estimates control finite-horizon sensitivity, and bounded acquisition and evaluation give a cardinality-independent transition-work contract. OPAL supplies a working persistent-state engine and controlled production measurements. After six successful million-node workloads, 18 matched fresh Release processes compared original and incremental receipt indexing at 10,000, 100,000 and 1,000,000 nodes. Degree, record shape, state dimension, active-target count and event schedule were fixed. Repaired resident-update P50 remained 0.083-0.084 microseconds; subsequent receipt-event P50 ranged from 4.458 to 6.459 microseconds across all scales. The original receipt path took 177.700-182.297 milliseconds at one million nodes. Reliability was tested alongside performance. Baseline and repaired engines produced byte-identical checkpoints, including continuation after fresh-process recovery. Replay, conflict, rejection, concurrent-event and allocation-failure checks passed; all 18 matched scaling processes passed their correctness gates. The repair preserves the state law and durable receipt format. These results connect the mathematical construction to a traced production transition with verified state continuity and measured scaling. Capacity-growth costs and formal scope are reported separately.

cs.OS · cs.DS

版本更新 · v2arXiv:2611.02650

ETH is Dead, Long Live SH

MIT Hardness Group, Josh Brunner, Erik D. Demaine, Fabian Frei, Timothy Gomez, Markus Hecher, Jeffery Li, Jayson Lynch

原文摘要

Among the recent flood of AI-based mathematical breakthroughs (the "mathpocalypse"), perhaps the biggest bombshells in theoretical computer science were two disproofs of the Exponential Time Hypothesis (ETH), via slightly subexponential algorithms for the classic $k$-Satisfiability problems. After the initial shock, the community quickly began to wonder at the implications: how do the hundreds if not thousands of ETH-based lower bounds in the literature need to change? It turns out, not much -- if we are willing to change the exponents ever so slightly. While Impagliazzo and Paturi's 1999 Exponential Time Hypothesis (ETH) is false, Stearns and Hunt's 1990 Satisfiability Hypothesis (SH) remains plausible. We prove that SH is equivalent to the recently proposed Weak ETH which rules out $O(2^{n^{1-\varepsilon}})$-time 3SAT algorithms for every constant $\varepsilon>0$, where $n$ is the number of variables. We show that SH (or Weak ETH) preserves the typical process for lower bounds: a polynomial-time reduction from a 3SAT formula of size $s$ (the number of variables plus clauses) to an instance of size $N\le s^{b+o(1)}$, for fixed blowup exponent $b>0$, rules out $O(2^{N^{1/b-\varepsilon}})$-time algorithms for every constant $\varepsilon>0$. By comparison, for an $N=O(s^b)$ blowup, ETH would have ruled out $2^{o(N^{1/b})}$-time algorithms. The framework also yields parameterized lower bounds ruling out a running time of $O(2^{k^{1/b-\varepsilon}})L^{O(1)}$ when the reduction produces an instance of length $L$ with parameter $k\le s^{b+o(1)}$. We also prove equivalence of the analogous counting hypotheses, #SH and Weak ETH, and that they rule out $O(2^{s^{1-\varepsilon}})$-time algorithms for #2SAT formulas of size $s$ for every constant $\varepsilon>0$.

cs.CC

新论文 · v1arXiv:2611.21905

Agile Monoped Hopping on Complex and Discrete Terrain via Kinodynamic Planning with Distilled MPPI Steering

Yue Qin, Denglin Cheng, Jingyu Liu, Yanran Ding

原文摘要

This paper presents FlowRRT, a kinodynamic planner for agile monoped hopping on discrete terrain. Steering for kinodynamic RRT on complex dynamical systems currently relies on motion primitives or closed-loop policies, limiting the diversity of motions or the planning speed. Distilling a trajectory optimizer avoids both, but existing methods regress a single solution per query, which fails on monoped hopping, where the feasible action set is narrow and disconnected. Our key insight is to treat steering as set-valued and predict the distribution of feasible actions instead. A multi-start MPPI expert collects the entire feasible set of each query offline, and conditional flow matching distills it into a generative sampler. With this sampler, FlowRRT coordinates successive jumps over long horizons and finds non-greedy maneuvers that are hard for reactive policies to discover. Benchmarks across different terrains show higher success rates and substantially shorter planning times than sampling-based baselines. On terrains with inclined platforms, FlowRRT reaches $80\%$ success within $1$~s, versus $24$~s for RL-RRT and $77$~s for SST. Sim-to-sim tracking in MuJoCo confirms accurate execution under full-order dynamics.

cs.RO

版本更新 · v2arXiv:2609.01852

The Memory Trust Gap: Reliance on Stale Evidence Across Model Scales in LLM Agents

Jundong Hu, Shekar Ramachandran

原文摘要

Persistent-memory agents can follow stale information despite current authoritative evidence. We study stale-evidence reliance across Qwen3 0.6/1.7/4/8B checkpoints using two controlled, action-scored suites: a Benefit suite requiring stored information and a Safety suite providing current evidence as static prompt text. We distinguish stale-value reliance from net accuracy loss relative to no memory. Benefit-suite reliance is 0.92-1.00 across checkpoints. In the Safety suite, larger checkpoints resist a plain stale note but incur substantial harm when it appears newer. A full $2\times2\times2\times2$ factorial manipulates label, recency, source authority, and position. Direct cross-size contrasts identify stronger recency effects in the larger tested checkpoints and a position-effect sign reversal within Qwen3. Metadata improves accuracy at all four checkpoints, with larger gains at 4B/8B; pre-resolution yields near-ceiling accuracy across all four. Llama-Instruct comparisons provide cross-family evidence for selected effects, while RGB and MisBench corroborate stale-value reliance. A framing control shows that the failure extends beyond memory-labeled content. These results characterize cue-dependent vulnerability across tested checkpoints without isolating capability or evaluating a complete memory lifecycle.

cs.AI · cs.CL

新论文 · v1arXiv:2611.21895

World-Grounded Reinforcement Learning for Pollution-Aware Bird Movement Modeling Using a Multi-Instance DQN Framework

Shivin Mangal, Divyansh Chawla, Anshu Garg, Isshaan Singh

原文摘要

Birds are modeled as environmentally sensitive agents whose movement reflects real-world air quality conditions, where AQI serves as a measurable world feedback signal representing pollution-driven consequences. Instead of relying on human preference or synthetic rewards, the system uses environmental outcomes as the primary training signal, enabling policy learning grounded in ecological impact. We propose a multi-instance reinforcement learning simulation in which agents navigate a 2D grid representation of India integrated with a precomputed AQI matrix and geographic state boundaries. The framework combines bird call metadata from the Xeno-canto dataset, spatial discretization, and real environmental data to learn adaptive navigation strategies under pollution constraints. To ensure scalability and efficiency, repeated real-time API queries are replaced by a structured AQI mapping that enables fast lookup during training and execution. Multiple baselines, including Monte Carlo, Q-Learning, and Value Iteration, are evaluated for comparison. Deep Q-Network demonstrates stronger stability, adaptability, and cumulative performance across varying environmental conditions, establishing a scalable foundation for world-feedback driven ecological decision modeling.

cs.LG

新论文 · v1arXiv:2611.21891

Call-Efficient ICD-10 Code Prediction via Supervised Candidate Generation and Selective LLM Invocation

Ilia Kopanichuk, Daniil Kartashev, Yuri Bolkonsky, Petr Anokhin, Vladimir Makharev, Iaroslav Bespalov, Dmitry V. Dylov

原文摘要

Automatic assignment of ICD-10 codes from short clinical diagnosis text is a practical medical text classification and normalization task. The task is difficult because diagnosis phrases are short and noisy, the ICD-10 label space is hierarchical, and the class distribution is strongly imbalanced. This paper presents a call-efficient ICD-10 code prediction pipeline that combines transformer embeddings, tree-based supervised candidate generation, candidate-level reranking, and a gated LLM assistance. The evaluation uses a leakage-safe validation protocol: model and threshold selection are based on out-of-fold cross-validation predictions, while final reporting is performed on a locked test set with no shared source-file groups. The supervised conditional parent-code candidate generator reaches 74.32% Top-1 accuracy on the locked test set and 93.05% Top-10 candidate coverage. A direct zero-shot GPT-5.6-Sol baseline reaches 81.01% Top-1 accuracy, but requires an invocation of this large model for every diagnosis. The final deployable system, Pre-LLM fusion, reaches 81.24% OOF-CV accuracy and 83.19% locked-test accuracy while calling the LLM for only 31.97% of cases. This improves the locked-test error rate from 25.68% for the supervised parent-code baseline to 16.81%, corresponding to a 34.5% relative error reduction. The results show that the most effective role of the LLM is controlled reranking on uncertain examples rather than unrestricted coding or full-dataset invocation.

cs.CL

版本更新 · v3arXiv:2506.03199

Quantum Cognition Machine Learning for Forecasting Chromosomal Instability

Giuseppe Di Caro, Vahagn Kirakosyan, Alexander G. Abanov, Jerome R. Busemeyer, Luca Candelori, Nadine Hartmann, Ernest T. Lam, Kharen Musaelian 等 14 位作者

原文摘要

The accurate prediction of chromosomal instability from the morphology of circulating tumor cells (CTCs) enables real-time detection of CTCs with high metastatic potential in the context of liquid biopsy diagnostics. However, it presents a significant challenge due to the high dimensionality and complexity of single-cell digital pathology data. Here, we introduce the application of Quantum Cognition Machine Learning (QCML), a quantum-inspired computational framework, to estimate morphology-predicted chromosomal instability in CTCs from patients with metastatic breast cancer. QCML leverages quantum mechanical principles to represent data as state vectors in a Hilbert space, enabling context-aware feature modeling, dimensionality reduction, and enhanced generalization without requiring curated feature selection. QCML outperforms conventional machine learning methods when tested on out of sample verification CTCs, achieving higher accuracy in identifying predicted large-scale state transitions (pLST) status from CTC-derived morphology features. These preliminary findings support the application of QCML as a novel machine learning tool with superior performance in high-dimensional, low-sample-size biomedical contexts. QCML enables the simulation of cognition-like learning for the identification of biologically meaningful prediction of chromosomal instability from CTC morphology, offering a novel tool for CTC classification in liquid biopsy.

q-bio.QM · cs.LG · quant-ph

版本更新 · v4arXiv:2610.22155

Realtime-Bench: Can Frontier Models Solve Real-time Interactive Problems?

Jingyu Tang, Xu Yang, Zhehao Zhang, Zheng Wang, Qi Cheng, Haozhe Zhao, Manling Li, Wenxuan Li 等 10 位作者

原文摘要

Real-time interaction requires agents to perceive evolving environments and act under temporal constraints, yet existing evaluations of real-time streaming models focus on conversational quality and streaming perception, leaving effective action under strict temporal constraints untested. We introduce Realtime-Bench, a suite of carefully curated dynamic tasks where delayed actions directly affect success. Under Wall Clock, where environments advance according to actual inference latency, the best success rate reaches only 44% and the best Pass^3 is only 36%, demonstrating that real-time interaction remains challenging. To separate serving latency from models' real-time streaming capability, we compare Wall Clock with Paused Clock, which removes inference delays as an idealized reference, and introduce Token Clock, which advances simulated time according to generated token counts, testing temporal awareness independently of physical serving speed. We find that temporal costs limit success in evolving environments and that higher reasoning effort has mixed effects under time pressure, reflecting a trade-off between decision quality and timely execution. Based on these results, we discuss future directions for improving real-time multimodal models, including adaptive reasoning with temporal awareness and continuous perception and action.

cs.AI

新论文 · v1arXiv:2611.21890

Domain-Tuned Diffusion Super-Resolution for 4K Facial Asset Maps

Elly Akhoundi, Mathieu Lamarre, Thibault Goyallon, Patrick Anderson, Étienne Danvoye

原文摘要

Physically based facial assets require 4K albedo and displacement maps for cinematic close-ups, yet generative pipelines commonly produce them at only 1K. Naively applying photographic super-resolution is a poor fit: facial UV maps have domain-specific statistics, and displacement is a scalar height field whose errors directly distort geometry. We present a domain-tuned diffusion upscaler for 4x (1024 to 4096) facial asset super-resolution. Starting from a bilinearly upsampled map, we adapt a frozen SDXL backbone per modality with lightweight LoRA modules, while a Tile ControlNet preserves input structure as the diffusion prior restores pore- and wrinkle-scale detail. On held-out maps, our method produces sharper, more faithful detail than bilinear and off-the-shelf GAN and diffusion super-resolvers, reducing albedo LPIPS from 0.21 to 0.08 while avoiding misplaced or excessive displacement relief. These results show that lightweight domain adaptation can produce render-ready 4K facial asset maps without a custom super-resolution architecture.

cs.CV · cs.AI · cs.GR

版本更新 · v2arXiv:2510.24653

PathoGaze1.0: An AI-Ready Navigation, Eye-Tracking, and Decision Dataset in Pathology

Veronica Thai, Rui Li, Meng Ling, Shuning Jiang, Jeremy Wolfe, Raghu Machiraju, Yan Hu, Zaibo Li 等 10 位作者

原文摘要

Performing cancer diagnosis on giga-pixel whole-slide images (WSIs) is an important but difficult visual search task. The large giga-pixel image size forces pathologists to navigate by zooming and panning to search tissue for malignancy. Diagnostic accuracy averages around 70%, and adding a second pathologist does not substantially improve consistencies between diagnoses. There is a lack of publicly available behavioral data to explain these diagnostic errors and inconsistencies. To fill this gap, we curated PathoGaze1.0, a comprehensive behavioral dataset capturing high-fidelity clinical diagnostic workflows across 11 different pathologist participants over two experiments 12-months apart (8 overlapping pathologists across both experiments), as they interpreted 397 WSIs from CAMELYON16, a set of publicly available breast cancer lymph node WSIs. The dataset comprises 18.69 hours of eye-tracking (171,909 fixations), and navigational scanning data along with 1,140 diagnostic decisions, all collected using a newly-developed testbed PTAH. The testbed supports interaction techniques found in clinical tools and collects viewport, mouse actions, and diagnostic decisions, enabling cross-links between WSI image stimulus, gaze, navigation behaviors, and diagnostic decisions. We share preregistered experimental materials, the datasets, and processing methods on the Open Science Framework (OSF).

cs.CV · cs.HC

新论文 · v1arXiv:2611.21888

A polynomial-time decoding algorithm up to half the square-root bound for quadratic residue and generalized quadratic residue codes

Phlippe Gaborit, Philippe Lebacque

原文摘要

Binary quadratic residue (QR) codes, introduced by Prange in 1958, are cyclic codes of prime length $p\equiv\pm1\pmod 8$ and dimension $(p+1)/2$ whose minimum distance satisfies the square-root bound $d\ge\sqrt p$. They include the Hamming code $[7,4,3]$ and the Golay code $[23,12,7]$, and were generalized by van Lint and MacWilliams to generalized quadratic residue (GQR) codes of prime-power length $q$, for which the bound $d\ge\sqrt q$ still holds and is attained when $q$ is a square. In coding theory, one usually first finds a lower bound on the minimum distance of a family of codes, and then a polynomial-time algorithm decoding up to half of this bound, as for BCH, Reed--Muller or Gabidulin codes. Until now, GQR codes had resisted this approach: no such algorithm was known, apart from individual short lengths. We give a deterministic algorithm which, for every odd prime power $q\equiv\pm1\pmod 8$, corrects every error pattern of weight at most $\lfloor(\sqrt q-1)/2\rfloor$ in the binary GQR code of length $q$, in $O(q^{9/2})$ operations in $\F_q$. We start from a description of QR codes due to Helleseth, which we extend to GQR codes, and in which the syndrome gives the values of the quadratic character of the error-locator polynomial. We then recover this polynomial by adapting to GQR codes the tools that Kopparty developed for recovering polynomials from noisy character values, with the decoding of dual BCH codes as an application. Since the square-root bound is attained when $q$ is a square, this radius is the largest that can be guaranteed uniformly on the class of GQR codes.

cs.IT

版本更新 · v2arXiv:2611.11010

Lightfoot: Graph Retrieval-Augmented Generation with Exact, Model-Free Document Deletion

Uday Allu

原文摘要

Graph-based retrieval-augmented generation (GraphRAG) improves multi-hop question answering, but deployed systems pay costs that research comparisons rarely measure: model calls at query time, building the graph, and removing documents correctly. We present Lightfoot, a GraphRAG system in which every entity, link, fact and synonym edge is a deterministic function of per-passage extractions that carry chunk-level provenance. Retrieval makes no language-model call, indexing makes one per passage, and deleting a document is a database transaction with no model calls: 14 seconds for 10% of an 11,402-passage corpus, after which the graph is identical to a fresh index of the surviving documents. On a corpus of near-duplicate names, HippoRAG 2's capped synonym lists leave 565 fewer edges than a rebuild, while Lightfoot remains exact. Under identical extraction, embedding and answer models on held-out questions of MuSiQue, 2WikiMultiHopQA and HotpotQA, Lightfoot is competitive with HippoRAG 2: with Holm-corrected paired tests it is better on MuSiQue at recall@5 and recall@10 (+4.8 and +5.5 points) and on 2WikiMultiHopQA at recall@10 (+3.4), and tied on HotpotQA, while HippoRAG 2 keeps a non-significant 3.0-point edge at 2WikiMultiHopQA recall@2. Seeding the walk from entities named in the question accounts for most of the 2WikiMultiHopQA gain. Answers tie on all three datasets, and the retrieval pattern holds with three larger embedders, up to Qwen3-Embedding-8B. The gain is not the reranker's: with the same cross-encoder, dense retrieval trails Lightfoot by 9.3 and 17.8 points recall@5 on the two multi-hop sets. Lightfoot also builds the cheapest index, 26-31% below HippoRAG 2; LightRAG needed 2,187 model calls and 3.7 hours for the same deletion, and fast-graphrag cannot delete.

cs.AI · cs.CL · cs.IR

新论文 · v1arXiv:2611.21882

WM-Quant: Rollout-Aware Post-Training Quantization for JEPA-Style World Models

William Hammer, Andrew Heuer, Nikhil Krishna, Matthew Grenier, Ramtin Zand

原文摘要

We present WM-Quant, to our knowledge, the first post-training quantization (PTQ) framework designed specifically for JEPA-style world models. Although world models typically feature relatively few parameters, action planning algorithms often require thousands of inferences, which can quickly become computationally expensive. PTQ is a natural way to reduce this cost. However, we show that in world models using autoregressive model predictive control (MPC) rollouts, quantization error can compound with each predictor step, resulting in action rankings from the quantized model becoming near-random. WM-Quant addresses this with three components that build on existing PTQ methods: early-encoder precision, control-path preservation, and rollout-aware scaling. Early-encoder precision protects accurate latent state and goal information to mitigate long-term error propagation. Control-path preservation allows the world model predictor to distinguish between fine-grained actions, significantly improving action rankings and execution. Rollout-aware scaling jointly calibrates encoder and per-step predictor activation ranges with respect to an entire rollout rather than a single prediction, reducing the cumulative rollout error. Across multiple world models, environments, and PTQ methods, WM-Quant substantially improves planning success at low bit-widths, approaching full-precision performance at W4A4.

cs.LG

新论文 · v1arXiv:2611.21880

Cancellation-Free Constrained SquareCB for Safe Controller Selection

Kartik Pandit, Sourav Ganguly, Arnob Ghosh

原文摘要

We study constrained contextual bandits, in which a learner chooses one of $K$ actions in each round to minimize loss while keeping the expected cost of its play below a budget. Existing reductions to online regression weight the cost by a Lagrange multiplier learned from past constraint residuals, so slack in some rounds can offset violations in others. We instead run SquareCB proposed in \citep{foster2020beyond} with a rectified weight on the cost: it is zero when the estimated cost is within the budget, and otherwise just large enough to bring it back to the budget. Under realizability and Slater's condition with margin $δ$, the algorithm attains regret $O(\sqrt{KT\,\mathrm{Reg}_{\mathsf{Sq}}(T)}/δ)$ and cancellation-free violation $\sum_t(V_c^{p_t}-b)^+=O(\sqrt{KT\,\mathrm{Reg}_{\mathsf{Sq}}(T)})$, where $\mathrm{Reg}_{\mathsf{Sq}}$ is the error of the regression oracle. The same bounds hold when the budget must be met in every context, potentially adversarially chosen contexts. In controller selection for a safe-control-gym quadrotor and a BOPTEST building, the violation is 6 to 9 times lower than that of primal-dual and virtual-queue baselines, with a lower or comparable regret.

cs.LG · cs.AI · stat.ML

新论文 · v1arXiv:2611.21879

Adaptive Memory for Online Reinforcement Learning under Sparse Model Changes

Sreejeet Maity, Sukanya Modak, Leonardo F. Toso

原文摘要

We study the discounted reinforcement learning (RL) problem in a setting where the underlying environment changes between deployments. Each deployment provides $H$ observations from a fixed MDP. The reward function and transition kernel are fixed within each deployment. Across deployments, at most an $ε>0$ fraction of the reward and transition kernel entries may change. In this setting, learning each model to deploy from scratch may discard useful observations from past deployments, whereas pooling all past observations can bias the estimates of entries that have already changed. We then propose an algorithm that for each state and action pair, selects how much history to reuse to estimate the reward and probabilities of the next state. We start with observations from the current deployment and add earlier deployment observations one at a time, the most recent first. After each addition, we check whether a single average reward agrees with the estimates from all histories considered until that point. We apply the same check to the distribution of next states. For each estimate, we leverage the last history that passes these aforementioned checks. The optimal control policy is then computed on top of the fitted MDP. We characterize the average gap between the optimal return and the return of the learner's policies over a finite number of deployments. We demonstrate that when $0< |S|ε\leq 2$, with $|S|$ being the cardinality of the state space, our algorithm needs only the observations from the most recent $O(1/ε)$ deployments, and achieves a long-run average gap of at most $\widetilde{O}(\sqrt{ε/H})$.

cs.LG · stat.ML

新论文 · v1arXiv:2611.21878

Knowing Which, Not How Much: Language Models Order Party Positions They Cannot Place

Yousef Attia, Jarosław A. Chudziak

原文摘要

Language models are increasingly asked where real actors stand: what a party's platform prioritises, what an institution has committed to. Such judgments are gradable, carrying both an order and a magnitude. Nearly every benchmark that tests them asks about one entity at a time, then reads that score as the model's knowledge of the domain. Order and magnitude, however, need not be acquired together. Text about social entities carries far more comparison than measurement, so a model may hold a dependable ordering with no anchored scale beneath it. We ask whether a model can place a documented social position on a scale or only order the actors it compares, and what that means for how such knowledge is measured. We test this on party positions in Japan and South Korea, where the Comparative Manifesto Project's coded data supplies a rare external answer key. Asked to order four parties, models beat the strongest fixed-answer baseline by more than a factor of two; asked to place one party on the same scale, they fall below it. The ordering survives inside those failed placements, and supplying the coded values lifts placement to near ceiling. Comparative and absolute questions about the same fact therefore measure different things here, and for party issue salience a benchmark built from one cannot speak for the other.

cs.CL · cs.CY

新论文 · v1arXiv:2611.21867

FaceRiff: Data-Efficient Unified Text-to-3D Face Asset Synthesis via Geometric Conditioning

Elly Akhoundi, Mathieu Lamarre, Thibault Goyallon, Patrick Anderson, Étienne Danvoye

原文摘要

High-resolution facial datasets with consistent UV-mapped geometry, albedo, and physically based displacement are scarce, yet such assets are essential for photorealistic digital-human rendering. Existing 3D face generation methods are often slow, single-modality, or require large specialized datasets. Meanwhile, large image diffusion models encode strong facial priors but do not natively produce lighting-independent textures, pore-scale displacement, or UV-aligned geometry. We present FaceRiff, a unified, data- and parameter-efficient framework for generating complete 3D face assets (mesh, albedo, and displacement) directly from text in a shared UV space. FaceRiff adapts a pretrained SDXL backbone with task-specific LoRA adapters and uses UV normal maps as geometric conditioning: per-instance normals guide albedo and displacement synthesis, while a dataset-mean normal provides a canonical spatial anchor for mesh generation. Using only 2,049 scans from approximately 350 identities and training on a single GPU, FaceRiff updates less than 3% of the backbone parameters per task and produces render-ready 3D face assets without per-instance optimization. Compared with FaceG2E and DreamFace, FaceRiff produces greater per-prompt diversity while preserving prompt alignment and fixed-topology UV consistency, and additionally generates lighting-independent albedo suitable for PBR relighting and pore-scale displacement absent from prior text-driven baselines. A blind study with 45 industry professionals shows that FaceRiff achieves substantially higher perceived realism than DreamFace, approaching that of real captures. Overall, FaceRiff demonstrates that pretrained diffusion priors can be adapted to unified 3D face asset synthesis under limited data.

cs.CV · cs.AI · cs.GR

新论文 · v1arXiv:2611.21865

PromptPack: Training-Free Augmentation of Diffusion Editors for Intent-Aligned Spatial and Semantic Control

Sophia Zalewski, Mustafa Munir, Enting Liu, Anjul Patney, David Dynerman

原文摘要

Many image-editing tasks require confining changes to intended spatial regions while exactly preserving user-designated source content. Instruction-driven editing pipelines often leave such user-intent underspecified, while inpainting editors condition generation on explicitly defined edit support but not on the spatial assignment of individual edit concepts. PromptPack addresses both limitations by equipping diffusion editors with two explicit 'Edit Guides': an edit-role partition that labels each pixel as 'Keep,' 'Permit,' or 'Modify' to specify where edits may occur, and a regional-prompt assignment that maps spatial regions to text prompts to specify which concepts apply where. These guides may be manually specified or, for source-object edits, automatically derived from a text-grounded SAM 3 mask. Without retraining, PromptPack injects independently embedded regional prompts into the editor's fixed-length text sequence and uses the Edit Guides to spatially mask attention, gate flow updates, and restore protected source pixels. We integrate PromptPack into FlowEdit FLUX, FlowEdit SD3, and FLUX Kontext and demonstrate improved requested-concept placement and protected-content preservation across all three integrations on PIE-Bench++. In a pairwise user study, PromptPack achieves an 87% preference rate over FlowEdit FLUX, establishing PromptPack as a practical, training-free interface for intent-aligned diffusion editing.

cs.LG · cs.AI · cs.CV

版本更新 · v12arXiv:2603.03339

Arapai: An Offline Adaptive AI Tutoring System for Competency-Based Learning in Low-Connectivity Environments

Joseph Walusimbi, Ann Move Oguti, Joshua Benjamin Ssentongo, Keith Ainebyona

原文摘要

Persistent learning deficits in low-resource educational environments stem from a structural mismatch: competency-based curricula demand individualized, iterative feedback, yet most digital learning tools require stable internet connectivity that is unavailable to millions of students. This paper presents Arapai, a fully offline adaptive AI tutoring system designed to close this gap through on-device intelligence. The system integrates a quantized lightweight language model with a four-stage pedagogical loop (diagnosis, instruction, worked example, and adaptive feedback) implemented without any external infrastructure. Following development, the system was demonstrated to a broad cross-section of stakeholders across multiple national academic and institutional exhibitions in Uganda, including technical judges, university faculty, tertiary educators, and secondary school practitioners. Expert observers provided informal feedback affirming Arapai's relevance and contextual fit, particularly with respect to Uganda's newly enacted Competency-Based Curriculum (CBC). These demonstrations establish a foundation of expert-observed validation; rigorous controlled efficacy studies with student cohorts are planned as the immediate next research phase. The system demonstrates that structured offline AI can function as viable educational infrastructure in low-connectivity emerging-market contexts, offering a practical pathway to AI-augmented learning without internet dependence.

cs.CY · cs.AR · cs.CL · cs.HC

新论文 · v1arXiv:2611.21859

Diversity Preservation Recovers Transcription Factor Motifs Across Challenging Confounder Landscapes

Alejandro Medina, Mary Lauren Benton

原文摘要

Motif discovery is a process where DNA sequences are searched for the short patterns, called motifs, that transcription factor (TF) proteins bind. However, the motif lengths are often unknown and motif attribution often competes with a near-identical motif of a related protein. In this work, we pair a motif genotype whose length emerges from an evolvable gate with MAP-Elites archives organized by three descriptor pairs, and validate every motif against the full HOCOMOCO database. On five liver ChIP-seq datasets spanning distinct confounder landscapes, a $(μ + λ)$ baseline differing only in survivor selection collapses onto a single motif and never ranks the target first across three datasets, while every MAP-Elites archive does so in at least 13 of 30 seeds on every dataset. Against the motif discovery tool MEME, run on the same sequences and validated identically, the archives recover the target signal as reliably on four datasets and more robustly when motif width is constrained. Neither method ranks the target above its closest relatives on HNF4A and JUND over the full width range, but narrow archive motifs rank HNF4A first in up to 43% of seeds, where MEME misses the motif in all settings. Diversity preservation improves our ability to recover consistent motif signal, although motif annotation remains a challenging problem for all methods we tested.

cs.NE · q-bio.GN

版本更新 · v3arXiv:2606.21869

The Language-Energy Divide: Measuring Energy Costs of Multilingual LLM Inference

Naihao Deng, Alissa Shen, Yiming Feng, Joan Nwatu, Jae-Won Chung, Mosharaf Chowdhury, Yulong Chen, Rada Mihalcea

原文摘要

Large language models (LLMs) are increasingly deployed in multilingual settings, yet the energy costs of serving these models across different languages remain poorly understood. We present a systematic study of inference energy consumption across languages with ML.Energy framework (Chung et al., 2026). We find striking disparities: energy consumption per output token varies by up to 8.3 times across languages, while total energy for a fixed set of requests varies by up to 179 times between the cheapest (English, 17.6 kJ) and the most expensive (Pashto, 3,147 kJ) languages. Our analysis shows that this disparity is driven by two compounding factors: (1) higher per-token energy costs for languages using complex or rare scripts, and (2) more tokens generated for low-resource languages. Moreover, we find a double cost + performance penalty: languages with the highest energy footprints also tend to achieve the lowest task accuracy. We reveal that the energy divide persists across models, hardware, and tasks, suggesting a systemic energy inequity in multilingual LLM deployment. Finally, we recommend that the community treat energy as a first-class evaluation axis, extend reporting checklists and model cards to include it, and adopt deployment-side mitigations for better energy efficiency.

cs.CL · cs.AI

新论文 · v1arXiv:2611.21855

Shortest-path percolation and structural robustness of quantum internet networks

Victor H. F. Francheto, Thomas Peron

原文摘要

We study shortest-path percolation on spatial networks that model the quantum internet based on optical fibers. In this process, the removal of paths describes the consumption of the entangled links used for communication. We compare finite and unrestricted path-length bounds for networks whose average degree either increases with system size or remains approximately constant. We show that networks become more robust as the average degree increases, whereas networks with approximately constant average degree exhibit a data collapse for the fixed finite bounds considered. Using Monte Carlo sampling of the shortest-path-length distribution, we estimate an equivalent path cost, a finite bound that approximately reproduces the degradation of the network in the unrestricted case. For networks with fixed average degree, this cost grows approximately as the square root of the number of nodes. Increasing the bound beyond this value requires additional computation time without appreciably changing the percolation dynamics. As a complementary result, for Erdős-Rényi random networks, we find an empirical relation between the initial average degree and the approximately common crossing point of the percolation curves for different path-length bounds.

physics.soc-ph · cond-mat.stat-mech · cs.SI · quant-ph