中 EN
Sign In Register

Academic Intelligence · Curated Daily

Explore the Frontier of Global Academia

AcademicHub aggregates real-time literature from top journals and preprint platforms. Build your personal research radar and let large language models compile cross-disciplinary analysis briefings automatically.

Sign In Register
02.
arXiv (CS.CL) 2026-06-12

MaxProof: Scaling Mathematical Proof with Generative-Verifier RL and Population-Level Test-Time Scaling

Authors:
Jiacheng Chen↗Xinyu Zhang↗Shunkai Zhang↗Yanmohan Wang↗Lin Li↗Tiancheng Qin↗Qin Wang↗Zhengmao Zhu↗Tianle Li↗Jingyang Li↗Zehan Li↗Binyang Jiang↗…

We present MaxProof, a population-level test-time scaling framework for competition-level mathematical proof in the MiniMax-M3 series. M3 first trains three proof-oriented capabilities – proof generation, proof verification, and critique-conditioned proof repair – using a defense-in-depth generative verifier engineered for low false-positive rate. These capabilities are merged into a single released M3 model. At test time, MaxProof treats the model as a generator, verifier, refiner, and ranker, searches over a population of candidate proofs, and returns one final proof through tournament selection. With MaxProof test-time scaling, the M3 model reaches 35/42 on IMO 2025 and 36/42 on USAMO 2026, exceeding the human gold-medal threshold on both.

Read & Discuss → View Source →
03.
arXiv (CS.LG) 2026-06-18

Task-Adaptive Parameter-Efficient Fine-Tuning for Weather Foundation Models

Authors:
Shilei Cao↗Hehai Lin↗Jiashun Cheng↗Yang Liu↗Guowen Li↗Xuehe Wang↗Juepeng Zheng↗Haoyuan Liang↗Meng Jin↗Chengwei Qin↗Hong Cheng↗Haohuan Fu↗…

arXiv:2509.22020v2 Announce Type: replace Abstract: While recent advances in machine learning have equipped Weather Foundation Models (WFMs) with substantial generalization capabilities across diverse downstream tasks, the escalating computational requirements associated with their expanding scale increasingly hinder practical deployment. Current Parameter-Efficient Fine-Tuning (PEFT) methods, designed for vision or language tasks, fail to address the unique challenges of weather downstream tasks, such as variable heterogeneity, resolution diversity, and spatiotemporal coverage variations, leading to suboptimal performance when applied to WFMs. To bridge this gap, we introduce WeatherPEFT, a novel PEFT framework for WFMs incorporating two synergistic innovations. First, during the forward pass, Task-Adaptive Dynamic Prompting (TADP) dynamically injects the embedding weights within the encoder to the input tokens of the pre-trained backbone via internal and external pattern extraction, enabling context-aware feature recalibration for specific downstream tasks. Furthermore, during backpropagation, Stochastic Fisher-Guided Adaptive Selection (SFAS) not only leverages Fisher information to identify and update the most task-critical parameters, thereby preserving invariant pre-trained knowledge, but also introduces randomness to stabilize the selection. We demonstrate the effectiveness and efficiency of WeatherPEFT on three downstream tasks, where existing PEFT methods show significant gaps versus Full-Tuning, and WeatherPEFT achieves performance parity with Full-Tuning using fewer trainable parameters. The code of this work is available at https://github.com/ShileiCao/WeatherPEFT.

Read & Discuss → View Source →
04.
arXiv (CS.CL) 2026-06-24

SciZoom: A Large-scale Benchmark for Hierarchical Scientific Summarization across the LLM Era

Authors:
Han Jang↗Junhyeok Lee↗Kyu Sung Choi↗

The explosive growth of AI research has created unprecedented information overload, increasing the demand for scientific summarization at multiple levels of granularity beyond traditional abstracts. While LLMs are increasingly adopted for summarization, existing benchmarks remain limited in scale, target only a single granularity, and predate the LLM era. Moreover, since the release of ChatGPT in November 2022, researchers have rapidly adopted LLMs for drafting manuscripts themselves, fundamentally transforming scientific writing, yet no resource exists to analyze how this writing has evolved. To bridge these gaps, we introduce SciZoom, a benchmark comprising 44,946 papers from four top-tier ML venues (NeurIPS, ICLR, ICML, EMNLP) spanning 2020 to 2025, explicitly stratified into Pre-LLM and Post-LLM eras. SciZoom provides three hierarchical summarization targets (Abstract, Contributions, and TL;DR) achieving compression ratios up to 600:1, enabling both multi-granularity summarization research and temporal mining of scientific writing patterns. Our linguistic analysis reveals striking shifts in phrase patterns (up to 10x for formulaic expressions) and rhetorical style (23% decline in hedging), suggesting that LLM-assisted writing produces more confident yet homogenized prose. SciZoom serves as both a challenging benchmark and a unique resource for mining the evolution of scientific discourse in the generative AI era. Our code and dataset are publicly available on GitHub (https://github.com/janghana/SciZoom) and Hugging Face (https://huggingface.co/datasets/hanjang/SciZoom), respectively.

Read & Discuss → View Source →
05.
arXiv (CS.CL) 2026-06-16

The Value Axis: Language Models Encode Whether They're on the Right Track

Authors:
Nick Jiang↗Isaac Kauvar↗Jack Lindsey↗

We investigate whether language models internally track the value of their current trajectory, defined as the likelihood that their ongoing strategy will achieve their goals. Using synthetic, in-context reinforcement learning data, we construct a "value" axis for Qwen3-8B. We find that activations along this axis distinguish between high vs. low verbalized confidence, rollouts without and with backtracking, and correct vs. corrupted code. Steering towards high value causally suppresses self-correction and reduces explanatory verbosity, while steering towards low value induces backtracking and exploration. We demonstrate that direct preference optimization (DPO) can increase the internal value of rewarded behaviors (e.g. use a certain word), causing the model to act more confidently after exhibiting them. Finally, we apply the value axis to study in-the-wild settings. For example, we find that Qwen assigns low value to politically sensitive chat queries after post-training and that supervised fine-tuning increases internal confidence within the training domain. Our results suggest that language models linearly encode an estimate of expected goal success that modulates their confidence in pursuing a direction.

Read & Discuss → View Source →
06.
medRxiv (Medicine) 2026-06-24

Cortisol Stress Response is Associated with Iron Status in Pregnancy

Authors:
Reid↗B. M↗Celestin↗G. F↗Georgieff↗M. K↗Mbayiwa↗Keenan↗

Background: Iron deficiency (ID) affects up to 40% of pregnant women in the third trimester, even in highly resourced and iron-supplemented populations, with adverse consequences for maternal health and long-term offspring development. Psychological stress may compromise iron status through hypothalamic-pituitary-adrenocortical (HPA) axis dysregulation and inflammation, but no study has directly examined cortisol in relation to iron status across human pregnancy. Objective: This longitudinal study examined associations between HPA function and maternal iron status across pregnancy and tested whether IL-6 and CRP mediated the relationship between cortisol and ferritin across gestation. Methods: One hundred sixty-eight pregnant Black women with Medicaid insurance completed up to four laboratory assessments across pregnancy. Salivary cortisol was measured before and in response to the Trier Social Stress Test, yielding basal and reactive cortisol indices. Serum ferritin, IL-6, and CRP were collected at each visit. Trimester-specific regression models examined cortisol reactivity in relation to ferritin; linear mixed-effects models with moderated mediation tested whether basal cortisol predicted ferritin via inflammation. Results: Higher cortisol reactivity was associated with lower ferritin specifically in the third trimester (std. {beta} = -0.197, p = .004). Higher basal cortisol predicted a steeper IL-6 rise across gestation (p = .002), and IL-6 was positively associated with ferritin (b = 0.236, p = .006), consistent with inflammatory iron sequestration. The indirect effect of basal cortisol on ferritin via IL-6 was statistically significant, and higher basal cortisol was negatively associated with cortisol reactivity in the third trimester. No pathway was observed through CRP. Conclusion: Greater cortisol reactivity predicted lower third-trimester ferritin, a pattern that suggests cumulative iron depletion, atypically sustained HPA reactivity in late pregnancy, or both. To our knowledge, this is the first prospective study linking cortisol reactivity to iron status across human pregnancy, identifying maternal stress physiology as a novel target for understanding and addressing gestational iron deficiency.

Read & Discuss → View Source →
07.
arXiv (CS.AI) 2026-06-12

Standardized Methods and Recommendations for Green Federated Learning

Authors:
Austin Tapp↗Holger R. Roth↗Ziyue Xu↗Abhijeet Parida↗Hareem Nisar↗Marius George Linguraru↗

arXiv:2602.00343v2 Announce Type: replace-cross Abstract: Federated learning (FL) enables collaborative model training over privacy-sensitive, distributed data, but its environmental impact is difficult to compare across studies due to inconsistent measurement boundaries and heterogeneous reporting. We present a practical carbon-accounting methodology for FL CO2e tracking using NVIDIA NVFlare and CodeCarbon for explicit, phase-aware tasks (initialization, per-round training, evaluation, and idle/coordination). To capture non-compute effects, we additionally estimate communication emissions from transmitted model-update sizes under a network-configurable energy model. We validate the proposed approach on two representative workloads: CIFAR-10 image classification and retinal optic disk segmentation. In CIFAR-10, controlled client-efficiency scenarios show that system-level slowdowns and coordination effects can contribute meaningfully to carbon footprint under an otherwise fixed FL protocol, increasing total CO2e by 8.34x (medium) and 21.73x (low) relative to the high-efficiency baseline. In retinal segmentation, swapping GPU tiers (H100 vs.\ V100) yields a consistent 1.7x runtime gap (290 vs. 503 minutes) while producing non-uniform changes in total energy and CO2e across sites, underscoring the need for per-site and per-round reporting. Overall, our results support a standardized carbon accounting method that acts as a prerequisite for reproducible 'green' FL evaluation. Our code is available at https://github.com/Pediatric-Accelerated-Intelligence-Lab/carbon_footprint.

Read & Discuss → View Source →
08.
arXiv (CS.CV) 2026-06-16

Mitigating Visual Hallucinations in Multimodal Systems through Retrieval-Augmented Reliability-Aware Inference

Authors:
Pratheswaran Hariharan↗Haiping Xu↗Donghui Yan↗

Multimodal large language models (MLLMs) have demonstrated strong capabilities in vision-language understanding and natural-language response generation. However, these systems can still produce overconfident predictions and hallucination-like outputs, particularly when the visual evidence is weak, ambiguous, or semantically inconsistent. Most existing approaches focus on improving multimodal representation alignment or retrieval-augmented generation, while providing limited mechanisms to quantify instance-level prediction reliability or identify incorrect visual outputs. This work proposes a retrieval-augmented reliability-aware inference framework for trustworthy multimodal visual understanding. The proposed framework constructs an external visual evidence database using pretrained visual embeddings and nearest-neighbor retrieval over normalized feature representations. Retrieved evidence is used to estimate prediction trustworthiness through multiple reliability indicators, including similarity strength, class-support agreement, evidence margin, entropy-based uncertainty, and an aggregate reliability score. Based on these signals, a decision gate determines whether the system should accept the prediction, answer with caution, or abstain/fallback when evidence is insufficient. A multimodal response-generation layer then produces a final user-facing response conditioned on the reliability decision. Experiments on ImageNet-100 demonstrate that the proposed reliability-aware framework improves accepted prediction accuracy from 85.84\% to 88.88\% at 89.04\% coverage. The hallucination-like accepted wrong-answer rate is reduced from 14.16\% to 11.12\%. These results show that integrating retrieval evidence, reliability estimation, and selective decision gating can improve calibration and reduce overconfident visual errors without retraining large multimodal models.

Read & Discuss → View Source →
09.
arXiv (CS.CL) 2026-06-25

Memory Makes the Difference: Evaluating How Different Memory Roles Shape Conversational Agents

Authors:
Yuxin Wang↗Paul Thomas↗Zhiwei Yu↗Yuan Gao↗Saeed Hassanpour↗Soroush Vosoughi↗Robert Sim↗Nick Craswell↗

Prior research on memory mechanism in RAG-based conversational system has emphasized how memory is stored and retrieved. However, far less is known about how memories with different functional roles influence response quality. Specifically, how they shape an agent's responses under varying conversational contexts and whether they lead to substantively different response behaviors. Existing evaluations in conversational system are also largely reference-based, insufficiently capturing the nuances in responses that may address users' preferences differently. In this work, we probe the impact of different memory types in shaping agents' responses. We present a fine-grained taxonomy of conversational memory, classify retrieved memories into different role types, and design a user-centric evaluation framework that simulates user perspectives. Through comparative experiments on long-term datasets and frontier LLMs, our analysis reveal many differentiated effects of memories: e.g., clarifying memory improves responses' factual accuracy and constraint awareness, making them more correct and personalized; irrelevant memory reduces topic relevance and degrades constraint awareness. Despite the power of frontier LLMs, these findings shed light on how different memory types can be leveraged to produce more personalized responses and inspire further research in this direction.

Read & Discuss → View Source →
10.
arXiv (CS.AI) 2026-06-12

Position: Generative Engine Optimization Creates Underexamined Risks, Governance Must Target Concentration, Disclosure, and Academic Blind Spots

Authors:
Yizhu Wen↗Nan Zhang↗Haohan Yuan↗Xun Chen↗Haopeng Zhang↗Hanqing Guo↗

arXiv:2606.12439v1 Announce Type: cross Abstract: Large language model (LLM) answer engines are increasingly used for information seeking, shifting visibility from ranked lists to synthesized answers. This enables Generative Engine Optimization (GEO), which targets LLM answer engines' evidence pool and generation. We analyze the search engine optimization (SEO) to GEO transition to identify two risks: (i) concentrated influence from low contestability and system sensitivity, and (ii) undisclosed commercial influence embedded in evidence and reasoning. We then formalize a general GEO pipeline to locate where optimization acts and compare academic and industry practices, revealing a third risk: (iii) academic-industry blind spots driven by visibility and evaluation asymmetries between offline setups and deployed systems. This position argues the need for answer-level governance and measurement: stronger contestability, high-precision disclosure, black-box auditing of material influence, and deployment-aligned metrics for exposure persistence.

Read & Discuss → View Source →
11.
arXiv (CS.CV) 2026-06-19

Collaborative Multi-Modal Coding for High-Quality 3D Generation

Authors:
Ziang Cao↗Zhaoxi Chen↗Liang Pan↗Ziwei Liu↗

3D content inherently encompasses multi-modal characteristics and can be projected into different modalities (e.g., RGB images, RGBD, and point clouds). Each modality exhibits distinct advantages in 3D asset modeling: RGB images contain vivid 3D textures, whereas point clouds define fine-grained 3D geometries. However, most existing 3D-native generative architectures either operate predominantly within single-modality paradigms-thus overlooking the complementary benefits of multi-modality data-or restrict themselves to 3D structures, thereby limiting the scope of available training datasets. To holistically harness multi-modalities for 3D modeling, we present TriMM, the first feed-forward 3D-native generative model that learns from basic multi-modalities (e.g., RGB, RGBD, and point cloud). Specifically, 1) TriMM first introduces collaborative multi-modal coding, which integrates modality-specific features while preserving their unique representational strengths. 2) Furthermore, auxiliary 2D and 3D supervision are introduced to raise the robustness and performance of multi-modal coding. 3) Based on the embedded multi-modal code, TriMM employs a triplane latent diffusion model to generate 3D assets of superior quality, enhancing both the texture and the geometric detail. Extensive experiments on multiple well-known datasets demonstrate that TriMM, by effectively leveraging multi-modality, achieves competitive performance with models trained on large-scale datasets, despite utilizing a small amount of training data. Furthermore, we conduct additional experiments on recent RGB-D datasets, verifying the feasibility of incorporating other multi-modal datasets into 3D generation.

Read & Discuss → View Source →
12.
arXiv (math.PR) 2026-06-16

Transposition Approach to Optimal Control of McKean-Vlasov SPDEs

Authors:
Liangying Chen↗Wilhelm Stannat↗

arXiv:2603.06245v2 Announce Type: replace Abstract: In this paper, we investigate an optimal control problem for McKean-Vlasov stochastic partial differential equations, in which the coefficients depend on the law of the state process. For systems with nonconvex control sets, we establish a Pontryagin-type stochastic maximum principle that provides necessary optimality conditions for admissible controls. The analysis is based on the classical spike variation method together with the introduction of an adjoint backward stochastic partial differential equation involving Lions derivatives with respect to probability measures. Our results extend the stochastic maximum principle for McKean-Vlasov controlled stochastic differential equations to the infinite-dimensional SPDE setting.

Read & Discuss → View Source →
13.
arXiv (CS.CL) 2026-06-24

AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages

Authors:
Hao Yu↗Tianyi Xu↗Michael A. Hedderich↗Wassim Hamidouche↗Syed Waqas Zamir↗David Ifeoluwa Adelani↗

Large language models (LLMs) are increasingly multilingual, yet open models continue to underperform relative to proprietary systems, with the gap most pronounced for African languages. Continued pre-training (CPT) offers a practical route to language adaptation, but improvements on demanding capabilities such as mathematical reasoning often remain limited. This limitation is driven in part by the uneven domain coverage and missing task-relevant knowledge that characterize many low-resource language corpora. We present \texttt{AfriqueLLM}, a suite of open LLMs adapted to 20 African languages through CPT on 26B tokens. We perform a comprehensive empirical study across five base models spanning sizes and architectures, including Llama 3.1, Gemma 3, and Qwen 3, and systematically analyze how CPT data composition shapes downstream performance. In particular, we vary mixtures that include math, code, and synthetic translated data, and evaluate the resulting models on a range of multilingual benchmarks. Our results identify data composition as the primary driver of CPT gains. Adding math, code, and synthetic translated data yields consistent improvements, including on reasoning-oriented evaluations. Within a fixed architecture, larger models typically improve performance, but architectural choices dominate scale when comparing across model families. Moreover, strong multilingual performance in the base model does not reliably predict post-CPT outcomes; robust architectures coupled with task-aligned data provide a more dependable recipe. Finally, our best models improve long-context performance, including document-level translation. Models and code have been released on [Huggingface](https://huggingface.co/collections/McGill-NLP/afriquellm) and [Github](https://github.com/McGill-NLP/AfriqueLLM).

Read & Discuss → View Source →
14.
arXiv (math.PR) 2026-06-11

The $K$-th nearest neighbor random walk on a Poisson point process gets trapped

Authors:
Anne-Laure Basdevant↗David Coupier↗Jean-Baptiste Gou\'er\'e↗Marie Th\'eret↗

arXiv:2606.11271v1 Announce Type: new Abstract: The $K$-th nearest neighbor random walk $(X_n)_{n \geq 0}$ on a homogeneous Poisson point process $\chi$ on $\R^d$ ($d\geq 1$), starts at the origin and at each step picks its next Poisson point among its closest neighbors according to i.i.d. labels having the same distribution as $K$. Our main result (Theorem 1) states that the number of Poisson points visited by $(X_n)_{n \geq 0}$ admits an exponential decay whenever the random variable $K$ has a bounded support (BS). In particular, the $K$-th nearest neighbor random walk visits finitely many Poisson points if and only if $K$ satisfies Assumption (BS). To prove it, we introduce the key notion of pioneer point which allows us to deal with the region of $\R^d$ already explored by $(X_n)_{n \geq 0}$. Still under Assumption (BS), we also prove an exponential decay for the Euclidean length of the trajectory performed by $(X_n)_{n \geq 0}$ (Theorem 2). Finally, and quite surprisingly, we exhibit an example of label distribution with bounded support for which the $K$-th nearest neighbor random walk discovers new Poisson points after a number of steps whose tail distribution is at least polynomial (Theorem 3).

Read & Discuss → View Source →
15.
arXiv (CS.LG) 2026-06-16

David vs. Goliath in Next Activity Prediction: Argmax vs. LSTM, Transformer, and LLM

Authors:
Hans Weytjens↗Ingo Weber↗

arXiv:2606.15868v1 Announce Type: new Abstract: Next activity prediction (NAP) is a cornerstone of predictive process monitoring (PPM), enabling organizations to move from retrospective analysis to proactive process steering. The PPM field has progressed from classical machine learning through deep learning architectures such as LSTMs and Transformers to large language models (LLMs). Despite growing model complexity, no benchmark jointly compares LLMs, Transformers, LSTMs, and simple baselines in a direct sequence modeling setting for NAP. In this paper, we fill this gap with a systematic benchmark. We compare vocabulary-adapted LLMs, Transformers trained from scratch, LLM-distilled Transformers, and LSTMs against a simple counting-based argmax baseline across seven real-life event logs. Our results tell a David vs. Goliath story: pretraining confers no consistent improvement over training from scratch, model size shows little effect on performance, and on most datasets the argmax baseline matches or approaches the performance of billion-parameter LLMs.

Read & Discuss → View Source →
16.
arXiv (CS.CL) 2026-06-16

SimSiam Naming Game: A Unified Approach for Emergent Communication and Representation Learning

Authors:
Nguyen Le Hoang↗Tadahiro Taniguchi↗Tianwei Fang↗Akira Taniguchi↗Masatoshi Nagano↗

Emergent Communication (EmCom) investigates how agents develop symbolic communication through interaction without predefined language. Recent frameworks, such as the Metropolis–Hastings Naming Game (MHNG), formulate EmCom as the learning of shared external representations negotiated through interaction under joint attention, without explicit success or reward feedback. However, MHNG relies on sampling-based updates that suffer from high rejection rates in high-dimensional perceptual spaces, making the learning process sample-inefficient for complex visual datasets. In this work, we propose the SimSiam Naming Game (SSNG), a feedback-free EmCom framework that replaces sampling-based updates with a symmetric, self-supervised representation alignment objective between autonomous agents. Building on a variational inference–based probabilistic interpretation of self-supervised learning, SSNG formulates symbol emergence as an alignment process between agents' latent representations mediated by message exchange. To enable end-to-end gradient-based optimization, discrete symbolic messages are learned via a Gumbel–Softmax relaxation, preserving the discrete nature of communication while maintaining differentiability. Experiments on CIFAR-10 and ImageNet-100 show that the emergent messages learned by SSNG achieve substantially higher linear-probe classification accuracy than those produced by referential games, reconstruction games, and MHNG. These results indicate that self-supervised representation alignment provides an effective mechanism for feedback-free EmCom in multi-agent systems.

Read & Discuss → View Source →
17.
Nature (Science) 2026-06-09

Good recycling starts at home — and benefits the world

Authors: Unknown Author

New research supports the value of household-level waste separation. But policies must also carefully consider consumer behaviours to maximize the quality of material collected. New research supports the value of household-level waste separation. But policies must also carefully consider consumer behaviours to maximize the quality of material collected.

Read & Discuss → View Source →
18.
medRxiv (Medicine) 2026-06-22

REPRODUCIBILITY OF 7T MRI MEASUREMENTS OF THE SUSCEPTIBILITY AND VOLUME OF HIPPOCAMPAL SUBFIELDS

Authors:
Adeyemi↗O. F↗Mougin↗Gowland↗P. A↗Rua↗Rodgers↗Hosseini↗A. A↗Bowtell↗

PURPOSE: The UK7T travelling head dataset was used to characterise the reproducibility of 7T measurements of the susceptibility of the hippocampal subfields, focusing on the Cornu Ammonis (CA1, CA2 and CA3), dentate gyrus (DG), subiculum (SUB), tail of the hippocampus (TAIL) and entorhinal cortex (ERC). METHODS: Susceptibility maps were created from whole-brain 3D single-echo GRE data (TE=20 ms; 0.7 mm isotropic resolution) using Multi-Scale Dipole Inversion. Automatic Segmentation of Hippocampal Subfields (ASHS) was applied to high resolution T1- and T2-weighted images for segmentation. The mean magnetic susceptibility and volume of hippocampal subfields was evaluated in 50 data sets, comprising 5 repeat acquisitions on 10 healthy participants (age 32 + or -6 years; 3 female). RESULTS: Averaging over subjects, susceptibility values spanned an 18ppb range over the hippocampus (ranging from -13.3ppb in DG to 4.7ppb in ERC). Susceptibility values in the larger hippocampal subfields showed a consistent pattern of variation across subjects, being generally more positive in ERC and SUB than in CA1 and more positive in CA1 than in DG and TAIL. The standard deviation of subfield susceptibilities over subjects ranged from 8.2ppb in the TAIL to 1.7ppb in CA1, and the average standard deviation across repeated measurements, which ranges from 1.7 to 4 ppb, was less than half of the inter-participant standard deviation in all subfields. Susceptibility values in the smaller subfields (CA2 and CA3) were more variable, but ICC(2,k) values for all subfields were >0.82. CONCLUSION: The reported data characterises the variation and reproducibility of hippocampal subfield susceptibility measurements at 7T.

Read & Discuss → View Source →
19.
arXiv (CS.LG) 2026-06-11

PAWS: Preference Learning with Advantage-Weighted Segments

Authors:
Aleksandar Taranovic↗Onur Celik↗Niklas Freymuth↗Ge Li↗Serge Thilges↗Huy Le↗Tai Hoang↗Rania Rayyes↗Gerhard Neumann↗

arXiv:2606.11982v1 Announce Type: new Abstract: Preference-based reinforcement learning (PbRL) learns policies from human trajectory-level comparisons, avoiding explicit reward design and expert demonstrations. Existing methods typically train utility functions on trajectory or segment-level preferences while relying on per-step utility estimates during policy optimization. This training and inference mismatch induces a distribution shift that severely degrades temporal credit assignment and limits policy learning. We analyze this issue and propose PAWS, a segment-based preference learning method that performs policy updates directly using segment-level advantage functions. By aligning utility training with policy optimization, PAWS preserves trajectory-level preference information and avoids unreliable per-step learning signals. Experiments on simulated robotic manipulation and locomotion tasks demonstrate that PAWS consistently outperforms existing PbRL approaches, highlighting the importance of distribution-consistent preference learning.

Read & Discuss → View Source →
20.
arXiv (math.PR) 2026-06-25

Higher moments of intrinsic volumes of random beta-prime polytopes

Authors:
Ferenc Fodor↗Bal\'azs Gr\"unfelder↗P\'eter Kevei↗

arXiv:2603.22224v2 Announce Type: replace-cross Abstract: We consider beta-prime polytopes, i.e., the convex hulls of iid random points chosen according to beta-prime distributions in $\mathbb{R}^d$. After suitable scaling, beta-prime polytopes converge in distribution to the convex hulls of Poisson point processes with power-law intensity functions. We prove moment convergence for the volume and all intrinsic volumes. Beta-prime polytopes are the push-forwards of spherical random polytopes on the upper open half-sphere of the unit sphere $S^d\subset \mathbb{R}^{d+1}$. We prove convergence of moments of the spherical volume difference of the half-sphere and the spherical random polytopes.

Read & Discuss → View Source →
21.
arXiv (CS.CL) 2026-06-16

Can LLM Coding Agents Reason About Time Series?

Authors:
Filip Rechtor\'ik↗Ond\v{r}ej Du\v{s}ek↗Zden\v{e}k Kasner↗

Large language models (LLMs) are increasingly being used for automated decision-making systems in finance, healthcare, or environmental monitoring. Time series data are ubiquitous in these fields, yet hard to process automatically. Can time series be analyzed by LLM agents? We examine three approaches: providing the agent with raw numerical data, using the LLM as a coding agent, or a combination of both. In the coding agent setup, the model iteratively queries the data using Python code. Using two time series understanding benchmarks, we show that agents with code access can outperform models processing raw data by up to 10%. However, even the best performing agent still answers about 22-34% of the questions incorrectly. To get insights into models' strategies and reasoning gaps, we analyze the model outputs with a strong LLM judge. Our analysis reveals that coding agents can select appropriate statistical tests, but often miss important nuances. Meanwhile, models with access to raw data can reach the right conclusions using back-of-the-envelope calculations.

Read & Discuss → View Source →
22.
arXiv (CS.LG) 2026-06-17

Stable and Steerable Sparse Autoencoders with Weight Regularization

Authors:
Piotr Jedryszek↗Oliver M. Crook↗

arXiv:2603.04198v2 Announce Type: replace-cross Abstract: Sparse autoencoders (SAEs) are widely used to extract human-interpretable features from neural network activations, but their learned features can vary substantially across random seeds and training choices. To improve stability, we studied weight regularization by adding L1 or L2 penalties on encoder and decoder weights, and evaluate how regularization interacts with common SAE training defaults. On MNIST, we observe that L2 weight regularization produces a core of highly aligned features and, when combined with tied initialization and unit-norm decoder constraints, it dramatically increases cross-seed feature consistency. For TopK SAEs trained on language model activations (Pythia-70M-deduped), adding a small L2 weight penalty increased the fraction of features shared across three random seeds and roughly doubles steering success rates, while leaving the mean of automated interpretability scores essentially unchanged. Finally, in the regularized setting, activation steering success becomes better predicted by auto-interpretability scores, suggesting that regularization can align text-based feature explanations with functional controllability.

Read & Discuss → View Source →
23.
arXiv (CS.CV) 2026-06-12

JointEdit3D: Feed-Forward 3D Scene Editing in a Unified Latent Space

Authors:
Xinnan Zhu↗Ruijie Xu↗Jiayu Ying↗Daoguo Dong↗Jiachen Xu↗Yuan Xie↗Xin Tan↗

Existing 3D scene editing methods typically rely on per-scene optimization over explicit 3D representations or cascaded edit-and-reconstruct pipelines, resulting in high test-time cost, limited 3D awareness, and structural inconsistencies. To couple appearance synthesis and geometry prediction during editing, we build on a unified RGB-geometry reconstruction-generation latent space and adapt it to feed-forward 3D scene editing. The resulting framework, JointEdit3D, performs asymmetric latent inpainting by observing only a single edited RGB reference latent and generating the remaining RGB views and edited geometry latent under source-scene anchoring. JointEdit3D introduces a dedicated SceneAnchor Branch to inject source-scene structure without forcing direct copying, and adopts edit/background-aware losses to balance edited-region fidelity with unedited-content preservation. To address the lack of paired resources for standardized 3D scene editing evaluation, we introduce SceneEdit3D-15K, a dataset with 15K paired editing samples and renderer-provided 3D annotations, together with SceneEdit3D-Bench, a curated 100-sample benchmark. Experiments show that JointEdit3D improves edited-region quality and 3D structural completeness over prior baselines while maintaining competitive background preservation.

Read & Discuss → View Source →
24.
arXiv (CS.CL) 2026-06-16

When the Same Musical Knowledge Forgets Differently: A Clean Probe of Pathway-Dependent Forgetting

Authors:
Yu Liu↗Zhiwei Yang↗Wenxiao Zhang↗Cong Cao↗Fangfang Yuan↗Kun Peng↗Haimei Qin↗Lei Jiang↗Jin B. Hong↗Hao Peng↗Yanbing Liu↗

A model can learn that the piano piece Für Elise is calm and reflective by listening to the audio or by reading a text description, but does it matter which route that knowledge took when it is later at risk of being forgotten? Forgetting research in multimodal models measures what knowledge is lost under adaptation, yet has not asked whether acquisition route affects how easily that knowledge is forgotten. We call this untested premise the Pathway-Invariant Assumption. Music understanding enables a clean test because a music clip and a canonical text description can be aligned to the same perceptual content, allowing the same knowledge unit to enter a model through listening or reading while the target remains fixed. Across multiple architecturally distinct audio-language models, we observe a consistent asymmetry: text-pathway knowledge is forgotten more than matched audio-pathway knowledge under identical adaptation pressure. To attribute this effect to route rather than confounds, we introduce the Paired Pathway Controlled Protocol (PPCP), a three-phase design that establishes matched pathway baselines, activates both pathways under symmetric supervision on the same knowledge pool, and applies identical forgetting pressure to both pathways. The gap is stable across models and gain-controlled analyses, persists when contradictory overwrite is replaced by correct-label cross-domain learning, remains under single-modality pressure, and is not removed by lightweight replay. Two independent routing-depth controls confirm that the effect is not explained by architectural depth, pointing to input representation as the dominant factor. Under PPCP, our results demonstrate that forgetting is highly route-dependent, establishing acquisition route as a new analytical dimension for forgetting research and multimodal system design.

Read & Discuss → View Source →
25.
arXiv (CS.CL) 2026-06-24

Towards Spec Learning: Inference-Time Alignment from Preference Pairs

Authors:
Dhriti Krishnan↗Tejas Goyal↗Jaromir Savelka↗

Steering a large language model (LLM) toward a desired behavior typically relies on an iterative process of hand-crafting a prompt based on a careful inspection of the model's responses. This is an involved, brittle, and error-prone process. Preference-based fine-tuning is a more rigorous but often prohibitively expensive solution. We propose spec learning, a framework that relies on a brief user instruction and a small set of preference judgments. These are compiled into specifications in the form of natural-language prompts for an LLM. Specifications condition LLMs at inference time, and no parameter updates to the underlying models are required. We show that the responses generated based on the compiled specifications often outperform direct preference optimization (DPO) on datasets from specialized domains whose preference signal is dense. Unlike opaque weight updates, the resulting specifications are human-readable and double as interpretable and transparent written embodiments of the preference signal that produced them.

Read & Discuss → View Source →