https://sspurlang.github.io/sspur/pulse/papers/The newest arXiv papers in cs.AI, cs.CL and cs.LG.enSat, 14 Nov 2026 01:32:20 +0000SSPUR Pulse https://arxiv.org/abs/2611.17334arxiv:2611.17334Thu, 12 Nov 2026 18:59:54 +0000Ellis Brown, Chris Hoang, Shengyi Qian et al.. How multimodal language models should encode visual input remains up for debate. Most models use a pretrained vision encoder, while some recent recipes train the encoder from scratch with the language model or remove it entirely. Vision encoders bundle together three design choices: a visual prior from upstream pretraining, dedicated vision capacity, and pre-fusion placement of that capacity (i.e… https://arxiv.org/abs/2611.17332arxiv:2611.17332Thu, 12 Nov 2026 18:59:52 +0000Jie Wang, Tong Zhu, JiaHao Kuang et al.. Model architecture shapes the scaling behavior, training efficiency, and inference cost of LLMs, making it a central focus of both industry and academia. In industry, teams such as those behind DeepSeek, Qwen, and Kimi routinely introduce architectural innovations, yet public reports rarely document which ideas succeeded, which were abandoned, and why. In academia, limited compute and inconsisten… https://arxiv.org/abs/2611.17329arxiv:2611.17329Thu, 12 Nov 2026 18:59:37 +0000Zeynep Yıldız, Satchit Chatterji, Ercan Atam et al.. Institutional portfolios operate under investment policy statements that specify asset allocation constraints, often conditioned on market regime information. Deep reinforcement learning has been shown to optimize financial performance but typically does not enforce rule compliance, making it unsuitable for real-world deployment. Existing neurosymbolic methods such as Probabilistic Logic Policy G… https://arxiv.org/abs/2611.17323arxiv:2611.17323Thu, 12 Nov 2026 18:59:06 +0000Ethan Lin, Akanksha Sarkar, Ziang Liu et al.. Applying machine learning systems to new domains introduces bottlenecks in both labeled data acquisition and manual pipeline engineering. While recent MLE agents show promise in automating pipeline design, they have been studied almost exclusively in fully supervised settings, leaving them untested in realistic regimes where labeled data is scarce but unlabeled data is abundant. In this work, we… https://arxiv.org/abs/2611.17320arxiv:2611.17320Thu, 12 Nov 2026 18:58:50 +0000Minhae Oh, Jeonghye Kim, Jinwoo Koo et al.. Remasking allows diffusion language models to revisit generated text, but effective correction requires knowing when and what to revise. Our diagnosis on Sudoku reveals two gaps. Hidden states contain error information that token confidence does not reliably expose, and remasking an isolated error usually reproduces it when the surrounding context remains fixed. These findings motivate DiVR , a v… https://arxiv.org/abs/2611.17316arxiv:2611.17316Thu, 12 Nov 2026 18:58:22 +0000Peiwen Li, Harshit Verma, Sophia Zhou et al.. Coding agents tackle successive tasks on an evolving repository, carrying forward edits and writing memory before the next task is known. Selecting memory for unknown next tasks requires balancing future usefulness against the costs of writing and reading it. The actor must also decide when acting on available memory saves effort as the task unfolds. We introduce Coding-Repository Memory (CoReMem… https://arxiv.org/abs/2611.17314arxiv:2611.17314Thu, 12 Nov 2026 18:57:49 +0000Abigail L. Tenenbaum, R. Thomas McCoy. Humans can learn rapidly from limited experience, while neural networks typically require extensive training. In this work, we investigate what properties of a neural network can enable it to learn rapidly. We study neural networks that have been optimized via meta-learning, a technique that exposes a network to many related functions to facilitate its subsequent learning of similar functions. Pr… https://arxiv.org/abs/2611.17313arxiv:2611.17313Thu, 12 Nov 2026 18:57:48 +0000Christopher Liaw, Aranyak Mehta, Grigoris Velegkas et al.. Context attribution asks how removing parts of a prompt changes a language model's output. We study leave-one-out (LOO) attribution for a prompt of $n$ tokens and a fixed measurement function. Computing all LOO values directly requires $n$ model evaluations. Under a stability assumption, our algorithm estimates all LOO values to error proportional to $\varepsilon$ with constant probability using… https://arxiv.org/abs/2611.17309arxiv:2611.17309Thu, 12 Nov 2026 18:56:56 +0000Benjamin Hoover, Hugh Blayney, James C. R. Whittington. Looped models have emerged as a prominent paradigm for reasoning, with successes on simple reasoning puzzles such as Sudoku-Extreme and Maze-Unique. However, these models can be unstable, slow to train, and use many parameters and FLOPs to solve simple puzzles. Descent on an energy not only mitigates these issues but also offers advantages: built-in early stopping, compute that adapts to problem… https://arxiv.org/abs/2611.17308arxiv:2611.17308Thu, 12 Nov 2026 18:56:54 +0000Félix Côté, Marie-Pier Côté. The short duration of general insurance contracts leads to rapid distributional shifts in the data, warranting regular model updates. This process may yield a post-update model that differs considerably from the pre-update model in both predictions and structure. Structural stability is upstream of predictive stability, yet existing structural stability metrics are mostly restricted to simple dec… https://arxiv.org/abs/2611.17303arxiv:2611.17303Thu, 12 Nov 2026 18:55:58 +0000Tosca Lechner, Vinayak Pathak, Sivan Sabato et al.. Debate has been recently proposed as a training protocol that is hoped to produce more aligned models even as model capabilities scale. The idea is that the human, instead of providing supervision in the form of labels, judges a debate between two models that compete to prove the correctness of opposite labels. It can be shown that functions of a much higher computational complexity can be correc… https://arxiv.org/abs/2611.17302arxiv:2611.17302Thu, 12 Nov 2026 18:55:56 +0000Mohamed Alimoussa, Amal El Kaid. Local Binary Patterns (LBP) remain among the most widely used texture descriptors, from industrial inspection to medical imaging. Yet their radius, sampling angles, color transform and thresholds are fixed by hand for each dataset. We present FLEX-LBP, which learns these parameters by backpropagation while every forward pass computes exact binary codes and hard histogram counts. A straight-throug… https://arxiv.org/abs/2611.17298arxiv:2611.17298Thu, 12 Nov 2026 18:55:09 +0000Gaurav Parmar, Zhengqi Li, Richard Zhang et al.. Interactive video-to-video applications require real-time generation while maintaining long-range temporal consistency. However, recent methods achieve speed by restricting attention to a fixed-length causal window. This limits long-range consistency whenever relevant history falls outside the context window. To address this limitation, we introduce vid2vid-long, a correspondence-based approach g… https://arxiv.org/abs/2611.17297arxiv:2611.17297Thu, 12 Nov 2026 18:55:06 +0000Arnab Sen Sharma, Eric Todd, Sheridan Feucht et al.. We investigate whether and how LLMs encode goals: internal neural representations that precede and guide action, and that are flexible enough to be realized through many different execution paths. We construct a coding setting with multiple sub-tasks that allows us to evaluate goal-seeking behavior and the selection of sub-goals. In this setting, we localize an encoding of the next sub-goal in th… https://arxiv.org/abs/2611.17294arxiv:2611.17294Thu, 12 Nov 2026 18:54:42 +0000Haozhe Wang, Lina Chen, Fanqi Wan et al.. On-device inference requires kernels that make effective use of local accelerators. Efficient training and evaluation of models that generate Apple Metal kernels benefit from native macOS tools and concurrent agent sessions. These sessions alternate between model requests, development commands, and GPU execution, leaving GPU capacity idle for much of a rollout. We present Metal Sandbox, infrastru… https://arxiv.org/abs/2611.17287arxiv:2611.17287Thu, 12 Nov 2026 18:52:33 +0000Jiacheng Xu, Shuo He, Fuxiang Zhang et al.. Reinforcement learning with verifiable rewards (RLVR) typically relies on a verifier that can reliably judge whether sampled solutions are correct. However, in many realistic settings, complete verification is generally unavailable, while cheap checks can only falsify failures. We formalize this regime as one-sided verifiability (OSV): verified negatives are reliable counterexamples, but verified… https://arxiv.org/abs/2611.17280arxiv:2611.17280Thu, 12 Nov 2026 18:51:05 +0000Gautam Galada, Supreeth Shastri, Sagnik Ray Choudhury. We propose Synthesis Smuggling, a novel threat model in which an adversarial LLM infers sensitive information about a user from routine assistance-seeking interactions and steganographically encodes that into innocuous-looking model-generated content. The user may then unknowingly transmit that information by sharing the generated artifact, say as an email, social-media post, or message, through… https://arxiv.org/abs/2611.17267arxiv:2611.17267Thu, 12 Nov 2026 18:45:58 +0000Pengcheng Zhou, Xuanyu Liu, Yanchen Yin et al.. We introduce GeoWorld, a controllable and continuously evolving geographic environment platform for LLM agents, and LiveGeo, a benchmark for evaluating reliable decision-making in real-world geographic environments using GeoWorld. Unlike existing evaluations that largely reason and plan over fixed geographic states, GeoWorld is grounded in real-world geographic structure and allows external world… https://arxiv.org/abs/2611.17246arxiv:2611.17246Thu, 12 Nov 2026 18:34:23 +0000Tiago de Souza Farias, Caio Moraes Porto, Matheus da Silva Fonseca et al.. Large language models (LLMs) have recently demonstrated the ability to generate parameters for variational quantum circuits without conventional optimization. However, the conditions under which this approach is effective remain unclear. Here, we investigate a frozen, general-purpose LLM that generates complete circuit parameter vectors from symbolic circuit and Hamiltonian descriptions, with opt… https://arxiv.org/abs/2611.17241arxiv:2611.17241Thu, 12 Nov 2026 18:28:08 +0000Samidha Verma, Ankit Anand, Sayan Ranu. Modern automated heuristic discovery (AHD) systems increasingly rival hand-crafted heuristics by leveraging Large Language Models (LLMs) to iteratively mutate and refine solutions. A natural assumption is that more powerful and more expensive models consistently yield better heuristics. We challenge this empirically: across multiple NP-hard combinatorial optimization problems, we find no consiste… https://arxiv.org/abs/2611.17239arxiv:2611.17239Thu, 12 Nov 2026 18:26:51 +0000Hannah Li. In interactive systems where users engage with LLMs, users are adaptive and, after observing LLM responses, can change their prompting strategy, as well as the type and difficulty of subsequent tasks. Thus model responses can influence the set of prompts on which the model is subsequently evaluated. We study the implications of this user adaptivity and prompt endogeneity for the evaluation of LLM… https://arxiv.org/abs/2611.17235arxiv:2611.17235Thu, 12 Nov 2026 18:23:56 +0000Elnaserledinellah Mahmoud Abdelwahab. This paper presents Al Ghasali, a computational framework designed to formalize and simulate Islamic jurisprudential (Fiqh) rulings. Building upon recent set theoretic formalizations of al sabr wa al Taqsim that extract minimal operational rules as simple Horn clause logical programs, this work proves that the semantics of closed Fiqh chapters collapse completely to deterministic Finite State Aut… https://arxiv.org/abs/2611.17230arxiv:2611.17230Thu, 12 Nov 2026 18:21:37 +0000Audrey Xie, Ludwig Schmidt, John Duchi. Many hypotheses compete to explain the successes of knowledge distillation. To help address this, we propose and analyze a mathematical model of distillation, which suggests that distillation's performance comes not from obtaining better models but from easier to optimize landscapes. For generalized linear models trained with stochastic gradient descent, we prove that distillation fits performant… https://arxiv.org/abs/2611.17228arxiv:2611.17228Thu, 12 Nov 2026 18:21:11 +0000Felix Zhou. We revisit the problem of estimating $k$ linear regressors with self-selection bias in $n$ dimensions up to error $\varepsilon$ with the maximum self-selection criterion, as introduced by Cherapanamjeri, Daskalakis, Ilyas, and Zampetakis [CDIZ23, STOC]. Our main result is a $\mathrm{poly}(n,k,1/\varepsilon)$ time algorithm for the unknown-index variant of this problem, thereby improving upon the… https://arxiv.org/abs/2611.17226arxiv:2611.17226Thu, 12 Nov 2026 18:20:22 +0000Ksenia Belkovskaya, Timur Lidzhiev, Maria Stebliy et al.. Graph-based collaborative filtering learns user and item representations by propagating embeddings over a user--item interaction graph. We investigate a modular, sheaf-inspired scoring extension placed on top of a three-layer LightGCN backbone. The motivation comes from cellular sheaves, where local vector spaces are related by linear restriction maps; here, this idea motivates structured, pair-d… https://arxiv.org/abs/2611.17225arxiv:2611.17225Thu, 12 Nov 2026 18:19:41 +0000S. J. Hamilton, A. Hauptmann, A. Hänninen et al.. This document describes open Electrical Impedance Tomography (EIT) measurement data that were collected from biomedical imaging inspired phantoms with moderate conductivity contrasts. The data were collected with the KIT4 and KIT5 EIT measurement devices at the University of Eastern Finland, and are intended for EIT image reconstruction algorithm development and testing. The data from thoracic im… https://arxiv.org/abs/2611.17223arxiv:2611.17223Thu, 12 Nov 2026 18:19:37 +0000Renato Paes Leme, Rad Niazadeh. We provide an efficient implementation of combinatorial Bernoulli factories. Such a factory turns coins of unknown biases $x$ into a random vertex of a polytope $P$ whose mean is exactly $x$. For every affine slice of the cube (the class of polytopes that admit a factory), we give a single algorithm whose expected number of coin flips and running time at every interior $x$ are polynomial in the d… https://arxiv.org/abs/2611.17220arxiv:2611.17220Thu, 12 Nov 2026 18:16:37 +0000Geoffrey Challen, Craig Zilles. Oral exams are compelling because (1) students can be asked to provide their reasoning for an answer and (2) their interactive nature permits follow-up questions to resolve ambiguous answers. But oral exams scale poorly. In this paper, we present scalable conversational assessment, an LLM-based system that conducts interactive text-based conversations with students to assess their ability to make… https://arxiv.org/abs/2611.17214arxiv:2611.17214Thu, 12 Nov 2026 18:10:46 +0000Aarush Kulkarni, Khang Nguyen, William Chang et al.. We study regret minimization in episodic Adversarial Markov Decision Processes (AMDPs) with unknown transitions and bandit feedback. In tabular AMDPs with $S$ states, $A$ actions, and horizon $H$, the best known $K$-episode regret is $\tilde O(HS\sqrt{AK})$, whose $\sqrt S$ gap to the $Ω(H\sqrt{SAK})$ lower bound has long remained open. In linear AMDPs with $d$-dimensional features, after a five-… https://arxiv.org/abs/2611.17212arxiv:2611.17212Thu, 12 Nov 2026 18:10:06 +0000Keshvi Tuteja, Yu-Hsiang Tsai, Gregor Olenik et al.. Deep neural networks for image processing are increasingly deployed on resource-constrained edge devices such as drones, medical systems, and smartphones, where limited computational resources and memory bandwidth pose significant challenges for efficient inference. Pruning during training often produces highly sparse convolutional kernels, offering the potential to significantly reduce computati… https://arxiv.org/abs/2611.17199arxiv:2611.17199Thu, 12 Nov 2026 18:02:53 +0000Soumi Sarkar, Subhadip Boral, Sarbani Palit et al.. Hyperspectral images (HSI) possess discriminating spectral characteristics with respect to land utilization due to high-dimensional information. Recently, unsupervised representation learning using graph neural networks has unveiled these utilization-specific complex spectral relations. This article has three main contributions. First, it proposes an index, the Mean Absolute Local Neighborhood In… https://arxiv.org/abs/2611.17196arxiv:2611.17196Thu, 12 Nov 2026 17:59:43 +0000Lucy Liu, Leon M. Kamp, Katia Bertoldi et al.. Automatic differentiation has enabled rapid progress on continuous-variable mechanical design problems, but gradient-based optimization still struggles with discrete variables. The standard approach is to optimize using a continuous relaxation of the choices, then commit to the nearest discrete design at the end. Why this procedure selects unfavorable basins is poorly understood. Here, we introdu… https://arxiv.org/abs/2611.17191arxiv:2611.17191Thu, 12 Nov 2026 17:55:33 +0000Tomohiro Sawada, Kartik Goyal. Low-Rank Adaptation (LoRA) often underperforms full fine-tuning, especially in heterogeneous and continual-learning settings in which substantial forgetting occurs. In general, it suffers from an undesirable learning-forgetting trade-off across diverse fine-tuning settings. We identify a potential contributor: LoRA adapters, especially high-rank ones, frequently fail to realize the available capa… https://arxiv.org/abs/2611.17186arxiv:2611.17186Thu, 12 Nov 2026 17:52:51 +0000Rui Sun, Jinfeng Peng, Zehua Cheng et al.. Federated learning (FL) enables independent data owners to train models collaboratively without centralizing raw data. However, identifying collaborators, preparing training tasks, and coordinating execution and incentives hinder its practical adoption by organizations without specialized expertise. We present Cortex-FL, an agent-native FL system that enables users to express learning objectives… https://arxiv.org/abs/2611.17169arxiv:2611.17169Thu, 12 Nov 2026 17:43:44 +0000Umar Ahmed, Maxim Dunaev, Mohammad Nour Al Awad et al.. A small fine-tuned parser that emits a well-formed but wrong tree fails silently: nothing downstream can tell. We ask whether the output format, a free design choice, controls how often this happens. On the PIZZA ordering benchmark we fine-tune two sub-billion models (Qwen3.5-0.8B and LFM2.5-350M) with LoRA on six lossless encodings of the same target tree, from 44.3 down to 10.9 tokens per order… https://arxiv.org/abs/2611.17167arxiv:2611.17167Thu, 12 Nov 2026 17:42:46 +0000Haochen Liu, Qianhao Zhu, Zixuan Wang et al.. Decision models answer a question with a typed output, such as a choice among options with a probability for each, so one model can choose an agent's actions, evaluate actions already taken, or score every action to teach a policy with reinforcement learning. Prior work has studied them mainly as actors, judging their usefulness by how well they act. We ask whether a decision model that acts poor… https://arxiv.org/abs/2611.17162arxiv:2611.17162Thu, 12 Nov 2026 17:39:35 +0000Anif N. Shikder, Ramit Dey, Jan Minac et al.. State space models (SSMs) are an emerging alternative to Transformers. SSMs offer linear-time inference and parallel training, but most implementations make the parallel scan efficient by making the recurrence diagonal: each channel of the hidden state is multiplied by an input-dependent real or complex number. These numbers commute, so the state records which inputs occurred but not their order.… https://arxiv.org/abs/2611.17161arxiv:2611.17161Thu, 12 Nov 2026 17:39:14 +0000Tanzeel Abbas, Tehseen Zia, David Windridge. Retrieval-augmented generation systems attach citations to attribute claims to their sources. These citations frequently misstate provenance, and standard verification cannot detect this. We measure provenance by counterfactual value substitution: a fact in the retrieved passages is replaced with a false variant and the model re-queried. We find that across five sizes of one model family and a se… https://arxiv.org/abs/2611.17152arxiv:2611.17152Thu, 12 Nov 2026 17:36:09 +0000Peter Sharpe, Carmelo Gonzales, Corey Adams et al.. Neural surrogates for computational fluid dynamics (CFD) are trained and ranked on datasets that each vary one design family in one flow domain. Such datasets cannot test the least that an aerodynamic foundation model must do: predict CFD-quality incompressible flow in seconds for new shapes in new enclosures. On DrivAerML, larger models and longer training barely lower Transolver's error; doubli… https://arxiv.org/abs/2611.17145arxiv:2611.17145Thu, 12 Nov 2026 17:31:50 +0000Yuanhong Wu, Mary Tsaryk, Sam Watson et al.. Understanding how visual attention influences human preference formation is a fundamental problem in consumer behavior and cognitive science. In this work, we investigate how eye movement patterns can be used to predict preference decisions when subjects compare two alternatives. To address this problem, we analyze an eye-tracking dataset, where subjects view pairs of face images and their gaze b… https://arxiv.org/abs/2611.17143arxiv:2611.17143Thu, 12 Nov 2026 17:30:34 +0000Carina Moreira Costa, Maria Eduarda Pinheiro, Martin Schmidt. We study data poisoning attacks on logistic regression with $\ell_2$-regularization, where an attacker chooses the features of a fixed number of injected observations with prescribed labels to maximize the logistic loss on the clean training data. Strong convexity of the defender's problem yields an exact reformulation through its first-order optimality conditions. Using convex conjugacy of the s… https://arxiv.org/abs/2611.17141arxiv:2611.17141Thu, 12 Nov 2026 17:30:03 +0000Peter Sharpe. A foundational principle of physics is that all coordinate systems are equally valid: the behavior of a PDE-governed system does not depend on the origin or orientation of the frame, nor on the units or mesh used to write it down, and neither should a surrogate model. Yet the neural surrogates with the lowest errors on today's aerodynamics benchmarks lack these invariances and fail under simple t… https://arxiv.org/abs/2611.17134arxiv:2611.17134Thu, 12 Nov 2026 17:26:01 +0000Fateme Golivand Darvishvand, Saurabh Mathur, Nikhilesh Prabhakar et al.. Managing children on ECMO (Extracorporeal Membrane Oxygenation) requires balancing short-term physiological stability with the long-term goal of preventing neurological injury. Standard reinforcement learning can propose risky, out-of-distribution actions, while imitation learning is limited to reproducing clinician behavior. We introduce Support-Constrained Fitted Q-Iteration (FQI-SC), an offlin… https://arxiv.org/abs/2611.17129arxiv:2611.17129Thu, 12 Nov 2026 17:22:17 +0000Ali Dadras, Alp Yurtsever. Federated learning experiments can produce reasonable-looking results even when the experiment implementation differs from the intended experimental protocol. Small errors in configuration, data partitioning, aggregation, randomness, evaluation, or checkpoint recovery may therefore affect a comparison without causing the program to fail. We present FedBrew, a research codebase for controlled and… https://arxiv.org/abs/2611.17128arxiv:2611.17128Thu, 12 Nov 2026 17:21:12 +0000Tanzeel Abbas, Ahmad Kamran Malik, Nafees Qamar et al.. Smaller language models (SLMs) make it feasible to deploy clinical natural language processing where large proprietary models are too costly, and parameter-efficient fine-tuning allows them to be specialised on commodity hardware. One obstacle persists after training: on medical tasks these models answer at far greater length than the clinical question requires. In clinical decision support and p… https://arxiv.org/abs/2611.17126arxiv:2611.17126Thu, 12 Nov 2026 17:20:49 +0000Tingfeng Li, Ming-Fang Huang, Shaobo Han et al.. Distributed fiber-optic sensing (DFOS) technologies such as Distributed Acoustic Sensing (DAS) generate continuous, high-resolution spatio-temporal data at rates exceeding hundreds of megabytes per second, creating severe challenges for real-time storage and transmission. While neural compression methods can achieve strong rate-distortion performance on sensing data, prior approaches either requi… https://arxiv.org/abs/2611.17121arxiv:2611.17121Thu, 12 Nov 2026 17:13:57 +0000Boris P. Belotserkovskii, R. Edward Watts. DNA-encoded libraries (DELs) are commonly used to find small-molecule binders to a target protein. But the hard part is no longer finding a hit. It is building a machine-learning model good enough to guide the medicinal chemistry that follows, one that can resolve the shift in affinity when a chemist changes a single atom. Such a model needs two things from its training data: dense sampling of th… https://arxiv.org/abs/2611.17116arxiv:2611.17116Thu, 12 Nov 2026 17:09:17 +0000Baptiste Debes, Tinne Tuytelaars. We introduce a mathematical framework for studying contraction of the distributional Bellman operator under Maximum Mean Discrepancy (MMD). We show that contraction reduces to conditional positive definiteness of an associated kernel constructed from the base kernel, which we call the Bellman difference kernel. Focusing on radial kernels, we use Schoenberg's theorem to connect this condition to B… https://arxiv.org/abs/2611.17115arxiv:2611.17115Thu, 12 Nov 2026 17:07:22 +0000Zimo Xu, Wenzhuo Liu, Hideki Nakayama. Enforcement defenses against prompt injection, such as CaMeL, Progent, IPIGuard and DRIFT, ask the agent's own model to write a security artifact (a program, a privilege policy, a dependency graph or a constraint list) that code then enforces. We ask whether small language model (SLM) agents can be trusted to write these artifacts. Across 15 open-weight models, 13 of them between 1.7B and 12B par… https://arxiv.org/abs/2611.17114arxiv:2611.17114Thu, 12 Nov 2026 17:06:54 +0000Petter Törnberg. As autonomous AI agents increasingly interact in persistent populations, their collective capabilities and alignment depend not only on individual models, but also on the institutions governing their interactions. Yet AI research has largely focused on individual agents. In this Perspective, we argue that institutional design must become a distinct focus of AI research. Drawing on institutional p…