Systolic Linear Attention: Co-Designing Tile-Grouped Gated Delta Attention for TPUs
Gagik Amirkhanyan, Nuojin Cheng
AI 速览：作者提出SLA，通过加宽键维度、分组衰减门和单核融合分块求解，使门控Delta线性注意力更适配TPU的宽脉动矩阵单元，并报告了相对KDA内核与预训练的训练加速。
Gated-delta linear attention layers such as Gated DeltaNet (GDN), Kimi Delta Attention (KDA), and GDN-2 are tuned for GPUs, where small tensor-core tiles and fast vector registers make per-channel decay rescaling and sequential in-block solves cheap. On Google TPUs, which concentrate compute in wide $256 \times 256$ systolic matrix units backed by a slower vector unit, porting these layers directly leaves half of each matrix tile empty and bottlenecks on vector operations. We present \textbf{Systolic Linear Attention (SLA)}, which reshapes gated delta attention around the matrix unit by widening keys to $256$ with grouped value heads, grouping decay gates into four $64$-channel tiles, and computing the chunk solve entirely with block matrix products in a single fused kernel. On TPU v7x, a full \texttt{SLA-R4} layer runs $1.6\times$--$2.5\times$ faster than the open-source Tokamax KDA kernel ($1.4\times$--$2.4\times$ with exact chunk inversion), driven primarily by head sharing and kernel fusion. Across a $2{,}048$-chip TPU v7x pre-training study of sparse Mixture-of-Experts (MoE) models from $6.5\text{B}$ to $63.1\text{B}$ parameters, SLA trains $11\%$--$24\%$ faster than KDA at similar perplexity, and reaches slightly lower perplexity than KDA in smaller matched-seed runs.
Marcin Moskalewicz, Anna Sterna, Karolina Drożdż, Kacper Dudzic, Marek Pokropski, Paula Flores
Building on a human-led thematic analysis of clinical life-story interviews (> 150,000 words) with inpatients with Borderline Personality Disorder, this study examines the capacity of large language models (OpenAI's GPT, Google's Gemini, and Anthropic's Claude) to support qualitative clinical analysis. The models' interpretative potential was evaluated using a mixed-methods approach. Study A involved blinded and non-blinded judges in phenomenology and clinical psychology. The experts assessed the validity of AI-generated content using semantic congruence, Jaccard coefficients, and multidimensional validity ratings, including credibility, coherence, the substantiveness of results, and grounding in qualitative data. In Study B, neural methods were used to embed human- and model-generated theme descriptions in a multidimensional vector space. This approach provided an objectified computational measure of the difference between human and model semantics and linguistic style. In Study C, complementary non-expert evaluations were conducted (N=115) to examine the influence of thematic verbosity on the perception of human authorship and content validity. Overall, the results of AI analysis showed a highly variable overlap (0-58%) with the human interpretation, while all models identified themes originally omitted by human researchers, proving their capacity to mitigate human bias. At the thematic level, external evaluators were unable to reliably distinguish human-authored themes from those generated by AI. In terms of content validity assessed against raw data by high-level experts in the blinded mode, the performance of Gemini 2.5 Pro was indistinguishable from that of humans. The comparison of semantic vector embeddings showed that its style was also the closest to humans.
A finite presentation < X | R > of a finite group is called `just finite' if removing any relation from R results in a presentation for an infinite group. It has been an open question (Kourovka Notebook, Problem 21.10) whether every finite group admits such a presentation. We resolve this conjecture in the affirmative.
We evaluate the extent to which agentic AI models can complete an end-to-end data science project, and how this improves with feedback. A 'writer' model with specific tools creates a complete data science project including downloading, cleaning, and analyzing a dataset, and writing a short paper about it. The project is executed and a 'grader' model evaluates it. The writer can then revise its submission, and this repeats for a total of three rounds. We consider seven models from five developers: Anthropic, OpenAI, Google, DeepSeek, and Z.AI. Each model is both a writer and a grader, so there are 49 writer-grader pairs and we run each pair three times. Of the 147 initial submissions, 107 run and render, and the average score increases from 59 per cent to 76 per cent after the final round of feedback. By writer the extent of the increase ranges from 3 percentage points for Anthropic's Opus 5.5 to 46 percentage points for Anthropic's Haiku 4.5, and by grader it ranges from 13 percentage points for Google's Gemini 3.8 Flash to 32 percentage points for Anthropic's Haiku 4.5. When a model grades its own writing the score tends to be about 9 percentage points higher than the average of what the other models give it. About half of the difference that we see from round to round is within the grader's own noise. Feedback generally results in the highest marks, but even just having the chance to revise, or having the additional budget, in terms of time or tool calls, without any feedback, results in considerable increases for some of the models but not for others. We find that how much such a loop helps depends on which model plays which role, but weaker models generally do not improve stronger models.
Mingxuan Li, Huanzhi Mao, Ruitao Zheng, Jarod Stanbury James, Yunze Li, Natanya Anderson, Joseph E. Gonzalez, Alex Dimakis
AI Agents are becoming digital coworkers that can browse the web, write code, run experiments, and assist humans over extended horizons. Yet existing benchmarks fall short of real knowledge work: many rely on synthetic data or isolated tasks, missing the complex, multi-party dynamics of collaborative projects, where critical information is scattered across conversations, code, and evolving decisions. Even simple workplace questions may require locating relevant channels and direct messages, recovering context from long discussions, inspecting repository state, reconciling stale or conflicting evidence, and synthesizing across sources. We introduce SlackBench, a benchmark for evaluating agents on realistic research and engineering workflows that require joint reasoning over workplace communication and evolving repository state. Grounded in privacy-preserving reenactments of real academic research projects, SlackBench provides reproducible environments with realistic multi-party dialogue, evolving decisions, and corresponding code changes grounded in GitHub repositories and Google Docs. It contains 140 queries spanning heterogeneous-source agentic search, abstention on unanswerable or false-premise questions, goal-directed summarization, and circumstance inference over long conversations. These tasks are not reducible to static corpus question answering: agents must determine which conversations matter, traverse threads, inspect repository state, and synthesize evidence across communication and code artifacts. Across 33 frontier models and agent harnesses, the best system scores around 80% and most agents score below 66%, highlighting substantial headroom for improvement. SlackBench thus provides an extensible framework for evaluating agents in collaborative work environments. The SlackBench dataset is available at https://huggingface.co/datasets/HuanzhiMao/SlackBench .
We study what LLMs do when a user applies pressure in an uncomfortable situation: a user insists, begs, flatters or grieves, and the model gives up a correct fact, writes a document it should refuse, or cheers a plan that will cost the user money, a multi-turn form of sycophancy. We send frozen multi-turn scenes to 60 models from 13 vendors and label each transcript with a trajectory (the model held its position or folded) and a manner (how it held or folded). Three scenes were used to build the codebook; six more were written afterwards and preregistered as a replication. Whether a model holds tracks its generation: fold rate correlates with a public capability index at Spearman -0.64 to -0.69 across the two scene sets, so newer models fold less, unlike earlier reports that sycophancy grows with model scale. How a model holds tracks the vendor: six of the 17 manner codes sort by vendor at permutation p <= 0.001, corrected across the codebook, and four do so again on the replication scenes. Anthropic's models name the feeling, warn and offer an alternative; OpenAI's do less of all three and almost never cite their own rules; Google's cite themselves and their AI nature when asked to do something; Meta's write the artifact they shouldn't. Six LLM coders from three vendors labeled every transcript, and three human coders labeled fifty held-out transcripts. The machines were the more consistent readers (Krippendorff's alpha 0.66 against 0.46), agreed with the author on trajectory at kappa 0.84 to 0.91, and matched the adjudicated reference at 0.83. For behavior a non-specialist can judge, the human contribution is authoring and bounding the codes and owning a small reference, not producing labels at volume.
Javeria Javed, Ahmed Salim, Muhammad Mahboob Ur Rahman
Accurate segmentation of the mitral valve from three-dimensional transesophageal echocardiography (3D TEE) is severely restricted by the scarcity of publicly available expert- annotated volumetric datasets. To address this critical limitation, we investigated Fourier Domain Adaptation (FDA), utilizing general-purpose generative-AI images (from platforms like Chat- GPT, Google Gemini, and Microsoft MAI) as annotation-free style donors for 3D mitral valve segmentation. We employed a DeepMitral residual U-Net model on the public MVSeg2023 dataset, comparing six conditions over three runs: real-only training, traditional augmentation, FDA with each generative source, and a non-AI FDA control using real-image style donors, plus an optimization-matched real-only control and a \b{eta} sensitivity analysis. Case-level paired testing showed that FDA with these generative sources led to significant reductions in boundary- and surface-related errors across all three donor conditions relative to the primary real-only baseline. However, the impact on the core segmentation metric, Generalized Dice Score (GDS), proved generator-dependent. ChatGPT-FDA showed a significant GDS advantage over the optimization-matched control but not over the primary baseline, and Gemini-FDA showed no significant GDS difference. Notably, only MAI-FDA provided a statistically robust and significant increase in GDS compared to the primary baseline. The \b{eta} sensitivity was not robust across training runs. These findings confirm that FDA remains an important augmentation mechanism. While generative-AI images offer a practical means of acquiring style diversity without new annotations, the non-AI control performed comparably to ChatGPT-FDA, so we do not conclude that this approach is superior to training with real-image donors.
Collaborative search enables users to address complex information needs jointly. Advances in large language models (LLMs) enable autonomous agents to support collaborative information seeking within shared communication spaces. However, little is known about how these agents shape search experiences, behaviors, and outcomes compared with conventional search tools. We conduct a user study with 80 students working in 40 pairs. Each pair completes four collaborative search tasks under four scenarios: QAAgent, which answers queries using an LLM's parametric knowledge; CoSearchAgent, which uses multi-user conversational context to reformulate queries, retrieve web information, and generate source-grounded answers, while asking clarifying questions when needed; an in-chat search plugin that returns result lists; and manual Google Search. After each task, participants complete several questionnaires assessing their search experience and independently write a report; after all tasks, they participate in individual interviews. Results show that LLM-based agents reduce perceived workload, support collaboration, and reshape communication and evidence-use practices. They also encourage more independent paraphrasing, albeit at the cost of report richness. To the best of our knowledge, this is the first systematic empirical comparison of LLM-based agents and conventional tools for collaborative search across search experience, user behavior, and collaborative outcomes.
Nishi Kanta Paul, Md Shihabul Islam Shovo, Audrija Chowdhury, Israt Jahan Esha
Automated diabetic retinopathy (DR) grading from colour fundus photographs is held back by two compounding challenges: severe class imbalance across five severity grades and the heavy computational cost of large scale architectures in resource constrained clinical settings. We introduce DR-LiteNet, a dual branch hybrid convolutional neural network that tackles both problems at once. The CNN branch extracts 1,280 dimensional deep features through EfficientNetB0, while the handcrafted (HC) branch produces a compact 51 dimensional lesion descriptor capturing vessel density, exudate intensity, texture entropy, and HSV colour histograms. Concatenating both branches yields a 1,331-d hybrid embedding, on which Adaptive SMOTE (ADASYN) is applied during training to synthesise minority class feature vectors, correcting distributional skew without touching the image space. A two phase training protocol first warms up the classification head on balanced features, then fine tunes the full network end to end. Grad CAM saliency maps deliver post-hoc clinical explainability. Tested on APTOS 2019 under five fold stratified cross validation, DR-LiteNet reaches 88.77% +/- 0.61 accuracy, a quadratic weighted kappa (QWK) of 0.884 +/- 0.010, and a macro F1 of 0.792 +/- 0.008 with only 4.69 M parameters beating ResNet50 (23.5 M) by DeltaQWK = 0.022 while staying practical on a single Google Colab T4 GPU. The source code of this project is available at: https://github.com/Nishi-Kanta-Paul/DR-LiteNet.
We compare how often two released language models give outdated numeric values from Norwegian statutes and regulations: NbAiLab/borealis2-26b-a4b-preview (Borealis 2), which the National Library of Norway further trained from google/gemma-4-26B-A4B-it (Gemma 4), and Gemma 4 itself. The study was preregistered (doi:10.5281/zenodo.23858965) and uses the closed norregel benchmark, scored deterministically against a value history. On the 411 of 503 primary items where both models gave an interpretable answer, Borealis 2 alone was outdated on 10 items and Gemma 4 alone on 11 (d = -0.2 percentage points, 95% cluster bootstrap interval [-2.7, +2.5], exact one-sided McNemar p = 0.668). By the preregistered rule this is no effect of interest: the whole interval lies within the smallest effect of interest of +/-10 percentage points. Among the secondary measures, only the share of answers giving a value that took effect after the question date differs after Holm correction; Gemma 4 gave such answers more often. An exploratory comparison of two smaller Borealis models had no discordant pairs and is not informative. The design compares released models and does not isolate the effect of any part of their training.
Andreas Schliebitz, Heiko Tapken, Martin Atzmueller
Self-supervised learning (SSL) typically relies on a backbone encoder followed by a small multilayer perceptron (MLP) projection head, which is conventionally discarded after training, while backbone quality is assessed via costly linear probing on labeled data. We argue that this approach including discarding the projector is rather computationally wasteful. Instead, we propose replacing the MLP head with a radial basis function network (RBFN), whose interpretable center and shape parameters can be exploited to judge representation quality without labels or a separate classifier. To this end, we introduce Scale-Normalized Separation (SNS), a novel label-free quality metric derived solely from the kernel centers and shapes learned during training. Across five canonical SSL architectures (MoCo, SimCLR, BYOL, SwAV and SimSiam) and four image classification datasets, we show that RBFN projection heads are competitive drop-in replacements for standard MLP projectors. We recommend constructing them with three RBF layers activated by the Gaussian radial basis function. Moreover, SNS exhibits strong to very strong positive correlation with established logistic regression metrics, demonstrating that a trained RBFN projector can act as a reliable proxy for backbone representation quality. We additionally publish a novel PyTorch compatible image classification dataset based on Google's Open Images V7 to facilitate reproducible research into representation learning.
Speech watermarking is becoming a key defense for digital copyright and content authentication. Existing attacks treat evasion as a removal problem: they erase the watermark, relying on detector access or on speech data that is hard to collect, such as paired clean-watermarked or clean target speaker speech. Commercial platforms are moving to close off these assumptions. TrustedAudio, for instance, restricts detection to authenticated users and denies meaningful detection feedback to unauthorized parties. Adversaries thus cannot tell whether a sample is watermarked or which watermark it carries, nor reliably remove it without degrading speech quality. In this work, we show that evading a speech watermark does not require erasing it: detection relies on the structural consistency of the embedded pattern, and disrupting this consistency is enough to defeat it. Building on this observation, we identify harmonic structure as an effective attack surface, as perturbing it breaks watermark consistency while preserving perceptual quality. We therefore propose ScrambleMark, a zero-query, watermark-agnostic, black-box attack that evades detection through harmonic scrambling. ScrambleMark combines a deterministic engine for discovering transferable harmonic perturbations with a generative model that learns randomized scrambling patterns, avoiding reliance on fixed attack rules. Across three commercial platforms (Google SynthID, Resemble AI, and TrustedAudio), ScrambleMark is the only attack to both evade detection and keep speech natural, achieving nearly 97% evasion and a 4.2/5 human rating, far above the 1.1 to 1.7 of prior attacks. Beyond evasion, ScrambleMark withstands adaptive recovery designed to reverse it, evidence that structural evasion is hard to mitigate in practice.
Yilun Hao, Ruixiao Yang, Hudson Hilal, Yongchao Chen, Kaizhi Qian, Yada Zhu, Yang Zhang, Chuchu Fan
Real-world problems span diverse domains, including program synthesis, symbolic reasoning, planning, and optimization, and often require fundamentally different solution paradigms. A central challenge for LLM-based reasoning is twofold: identifying an appropriate problem-solving approach and formulating it correctly. In this paper, we propose AlgoPilot, a cross-paradigm reasoning framework that enables adaptive selection of the solving approaches and structured problem formulation prior to execution. The AlgoPilot framework contains three components: 1) a SteerLM, trained via supervised fine-tuning (SFT) and reinforcement learning (RL), that adaptively selects the solving strategy (paradigm and modes) based on input problems, 2) the corresponding expert GuideLM, trained via SFT, that generates structured formulation guidance tailored to the selected strategy, and 3) an ExecutionLM that generates solutions conditioned on both the selected strategy and its formulation. Unlike prompting-based methods that rely on fixed reasoning templates, our approach learns to adaptively choose and structure problem-solving strategies based on the input. We build AlgoPilot on Qwen3-8B and evaluate across 5 problem categories and 13 benchmarks with 66 unique domains, comparing against both the latest similarly scaled small models and large LLMs. AlgoPilot significantly improves average accuracy for the same model from 38.4% to 72.9%, outperforming the best large LLM prompting baseline GLM-5 by 11.7%. Ablation studies confirm the effectiveness of each component. Notably, we show that the learned steering and formulation modules can transfer across execution models, suggesting their potential to improve other LLMs. Project page: https://sites.google.com/view/algopilot.
Variable fonts enable continuous variation of glyph geometry along semantic design axes such as weight, width, slant, and optical size. However, constructing a variable font from a static font remains a labor-intensive process requiring expert typographic design and manual specification of glyph variation data. We introduce NIV (Neural Axis Variations), a method that automatically converts a static font into a fully functional variable font. Given glyph outlines and a set of desired design axes, NIV predicts per-point displacements. The model operates directly on vector glyph geometry and employs a novel Property Embedding mechanism that captures interactions between multiple axes, enabling consistent multi-axis variation within a unified framework. We train NIV on a newly constructed dataset derived from variable Google Fonts, comprising over one million variation tuples. The resulting model generalizes across unseen code points, unseen font styles, high-complexity CJK glyphs, and even out-of-distribution handwriting inputs. The generated outputs are standard variable font files supporting continuous interpolation via existing rendering engines. To facilitate research, we release the dataset, the complete training and inference implementation, and trained models at https://github.com/ndvbd/NIV. Beyond typography, our approach demonstrates how structured geometric objects with continuous parametric variation can be synthesized using neural deformations.
Danyal Sarfraz, Raja Mubashar Karim, KwanMyung Kim
Recent advances in generative AI have enabled multiple pathways for high-fidelity video synthesis: text-to-video, image-to-video animation, and video outpainting. Empirical side-by-side evaluations of how untrained viewers perceive and distinguish these outputs from real footage remain scarce. In this study, we systematically compare human detection accuracy across these three AI generation techniques within three thematic contexts: historical footage, film and media content, and natural environments. We constructed a balanced stimulus set comprising equal amounts of real and AI-generated videos (18 each). The AI clips were evenly distributed across the three generation methods using Google's Veo 3, Lightricks LTX Video, and Wan VACE. All videos were produced and standardized within the ComfyUI framework to ensure consistent quality and duration. 87 participants judged every clip in a binary forced-choice task ("real" vs. "AI-generated"). Participants correctly identified videos 60% of the time on average. Image-to-video clips were recognized most accurately (79%), followed by real footage (64%), outpainting (49%), and text-to-video (43%). Accuracy also varied by theme: film and historical scenes yielded higher detection rates than environmental clips, which were frequently mistaken for AI. Logistic regression confirmed significant effects of both technique and theme as well as their interaction (p < 0.001), indicating that detection success depends jointly on how the content was generated and what it depicts. Findings reveal a consistent bias toward assuming synthetic origins and highlight that perceptual realism in AI video is shaped more by context than by model type, underscoring the importance of media-literacy approaches and context-aware evaluation tools for navigating increasingly synthetic visual media.
Is It the Curriculum or Just the Easy Start? Benchmarking Unsupervised Environment Design on Quadrupeds
Jakub Gilewicz, Dominik Galus, Piotr Wojdas
AI 速览：该论文在五种3D四足机器人上统一基准测试了十四种无监督环境设计、域随机化和机器人课程方法，发现没有一种方法在所有机器人和关卡集上均最优，且简单或低难度训练关卡的方法在分布内随机关卡上表现突出。
Unsupervised environment design (UED) methods such as PLR, ACCEL, SFL and PAIRED are compared mostly on 2D mazes, grid worlds and BipedalWalker, while terrain curricula for legged robots are hand-designed and evaluated against a few baselines. We present a benchmark that runs fourteen UED, domain-randomisation and robotics curricula under one protocol on five 3D quadrupeds in MuJoCo MJX: the closed-chain W01-TEK and the open-chain Unitree Go1 and Go2, Google Barkour vB and Boston Dynamics Spot. Levels are vectors of eight interpretable terrain parameters, and all curricula drive the same PPO student at matched environment-step budgets. The evaluation reports success on random and held-out structured terrains, interquartile means with stratified bootstrap intervals, capability envelopes and a level-by-level domination comparison. We compare methods within each robot and report, per level set, the top group of methods not significantly below the best. No method is in the top group on every robot and level set. Methods with easy or low-difficulty training levels, curated or not, fill the top groups on in-distribution random levels of Go1 and Barkour, and of W01-TEK apart from PAIRED+BiBC. The success-driven terrain curriculum from legged-robot practice is in a top group most often, domain randomisation on uniform levels in none, and PAIRED in none outside Spot, where the fourteen methods lie within 3.8 points on random levels of difficulty 0.25. On three robots the leader on in-distribution random levels is not the leader on held-out terrain. We collect the results into a per-robot guide to choosing a curriculum.
TorchTPU - A Hybrid Eager/Compiled Architecture for Native PyTorch Execution on TPU
Claudio Basile, Michael Voznesensky, Gunhyun Park, Ben Wilson, Parker Holloway, Zhanyong Wan, Vlad Belous, Manish Khanna 等 53 位作者
AI 速览：Google 与 Meta 合作开源 TorchTPU，为 Google TPU 提供原生 PyTorch 后端，兼顾开箱即用与训练/推理性能。
TorchTPU is an open-source collaboration between Google and Meta that introduces a native and principled PyTorch backend for Google TPUs. It enables seamless out-of-the-box model execution and TPU migration while delivering industry-competitive inference and training performance through eager execution optimizations, torch.compile integration, and native hardware acceleration.
Subhashini Venugopalan, Jenny Tan, Taylor Roper, Jimmy Tobin, Anton Kast, Alicia Martin, Sam Sepah, Amy Pavel
Audio descriptions (AD) are essential for making visual media accessible to blind and low-vision (BLV) users. While Multimodal Large Language Models (MLLMs) offer a scalable solution for generating audio descriptions on-demand, their performance on ``inline'' audio descriptions -- which must fit within existing silences in a video -- remains under-explored, particularly for diverse user-generated content (UGC). We present a comprehensive investigation into MLLM-generated inline audio descriptions. First, we introduce MTG-AD, a professionally annotated dataset of 36 hours of videos, each averaging $\sim$5 mins. and 8k+ descriptions in total, co-created by paired blind and sighted AD professionals to provide a high-quality benchmark for this task. Second, we propose MTG-AD-Eval, an evaluation framework based on audio description expert guidelines that provides fine-grained, reference-free actionable metrics for model improvement. Crucially, our framework evaluates the end-to-end task: it renders the generated scripts into audio tracks to assess their fit within the original video's natural silences, minimizing disruptive overlaps. Our analysis across frontier MLLMs reveals that while models accurately describe visual events and substantially reduce speech overlaps, gaps persist in narrative pacing and flow. Finally, we empirically validate that our automated and reference-free metrics correlate significantly with human quality ratings from sighted raters and 7 BLV participants, finding that while users narrowly prefer human-authored audio descriptions, they find MLLM-generated audio descriptions highly beneficial for comprehension. We provide actionable system design guidance and release MTG-AD at https://hf.co/datasets/google/mind-the-gap-inline-ad and MTG-AD-Eval at https://github.com/google/mtg-ad.
Analogical reasoning is widely considered a fundamental component of human intelligence. Few benchmarks are designed to explicitly test far analogical reasoning - reasoning between domains that are superficially dissimilar yet share deep relational or structural similarities. In this work, we introduce an analogical reasoning benchmark called Abstraction and Reasoning Corpus of ANAlogy (ARCANA). The tasks of ARCANA require performing far analogical mappings from real-world objects or physical phenomena to abstract grid-based transformations, different from the well-known ARC-AGI benchmark which primarily emphasized innate human priors. We validate ARCANA through a human study, demonstrating that humans solve these tasks reliably. In contrast, evaluations across a range of large language models show a distinct performance gap compared to humans, revealing systematic failure modes that remain challenging for current models. To address this gap, we propose ConceptSIR, a reasoning framework based on concept slippage, introspection, and reflection. Experiments demonstrate that ConceptSIR enables standard LLMs to solve several challenging ARCANA tasks that were unsolved by direct prompting in our evaluation. Overall, our results suggest that far analogical reasoning remains an open problem for contemporary LLMs. All the materials are available at sites.google.com/view/arcanacogtest. The complete codebase, including benchmark, and ConceptSIR, is publicly available at https://github.com/density-11/arcana-test. All ARCANA data and task instances can be accessed and explored at https://arcanacogtest.org/.
Sam Agnoli, Brigid Reilly, Lily Barna, Emőke-Ágnes Horvát
Google's AI Overviews (AIOs) are the first result users encounter when searching the web, shifting the work of selecting and evaluating sources from the user to Gemini. This may be especially consequential for health information, given that Google's traditional algorithm was designed to prioritize quality sources, and inaccurate information carries health risks. We evaluated 3,611 benchmark health questions, comparing each AIO against that question's top five traditional results on source selection and quality. We also evaluated the AIOs' statement grounding in their cited sources and how accurate AIOs were against physician-written rubrics. Overviews appeared for 96.4% of questions but drew on substantially different evidence than traditional search: only 44.4% of the top five results were cited by the corresponding AIO, which instead overrepresented commercial health publishers (e.g., WebMD, Healthline) at the expense of academic and government sources (e.g., Cleveland Clinic, National Health Service). AIO-cited pages scored lower than traditional results on four of six well-established credibility criteria. The grounding of AIO statements to their sources was poor: after correcting for classifier error, an estimated 56.7% of cited statements were supported by their cited page, and a further 20.8% of all statements were factual claims with no citation at all. When AIOs were graded against physicians' rubrics, they earned 38.2% of all available points (e.g., accuracy 54.2%, completeness 21.6%), and among the subset of AIOs whose physician rubric anticipated a specific inaccurate or potentially harmful claim, 22.2% of AIOs contained the harmful content. Notably, the accuracy of AIOs was unrelated to the grounding of their statements. We recommend AIO literacy interventions for users and holding AIOs at least to Google's own health-ranking standards.
Henrique Schechter Vera, Sahil Dua, Jovan Andonov, Sara Smoot, Sebastian Russo, Ivan Llanos, Max Matveev, Jae Hyeon Yoo 等 73 位作者
We present EmbeddingGemma 2, an open multimodal embedding model which natively represents text, images, audio, video, and combinations thereof in a unified space. Our approach introduces improvements such as simplified per-layer embeddings and a multi-stage curriculum with modality and objective progression which ensure strong quality across unimodal, cross-modal, and mixed-modal evaluations. Notably, in comprehensive embedding benchmarks across all modalities, EmbeddingGemma 2 establishes a new Pareto frontier, outperforming both specialist models and models two to twelve times larger. Furthermore, our design offers flexibility for resource-constrained environments, supporting selective modality encoder loading, adjustable vision token budgets, truncatable embedding dimensions, and quantized weights. We present ablations analyzing key decisions and release EmbeddingGemma 2 to promote further research.
Alejandra Zambrano, Sara Vera Marjanovic, Imene Kerboua, Xing Han Lù, Leila Kosseim
Despite recent advances, LLM-based web agents still struggle with limited exploration, omission of critical steps, and sensitivity to task constraints. Prior work suggests that many of these failures stem from weaknesses in planning, yet the impact of alternative natural language plan representations remains unexplored. To address this, we introduce PlanAhead, a static planner-executor framework that evaluates the impact of plan representation on agent performance. We first automatically categorize WebArena tasks into 3 difficulty levels, enabling consistent difficulty grading without human annotation. Then we systematically evaluate 4 different plan representations on the tasks categorized as hard: sequential subgoals, narrative, pseudocode, and checklist; across different families of multimodal LLM-powered agents(OpenAI, Alibaba, and Google). To account for stochastic variability, we introduce a novel decomposition of agent Success Rate (SR) into 2 interpretable evaluation metrics: Achievement Rate (AR) and Solved-Task Consistency (STC). Our results show that both, the plan formulation and the underlying LLM generating the plan, significantly influence web-agent robustness and task success.
Tosin Adewumi, Martin Karlsson, Marcus Liwicki, Mikael Sjödahl, Lama Alkhaled, Rihab Gargouri, Nudrat Habib
This work presents a systematic review of the application of natural language processing (NLP) along the entire battery life cycle, instead of the review of one stage or method, and introduces a novel conceptual technical language processing (TLP) framework for the European Union (EU)'s proposed digital battery passport (DBP) and other general battery predictions, as the main objective. It follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method and employs three reputable databases or search engines, including Google Scholar, Institute of Electrical and Electronics Engineers Xplore (IEEE Xplore), and Scopus, searched in October, 2025. The search criteria are 'text', 'language processing', and 'battery' while the exclusion criteria are non-relevant papers and those published before 2017, which was applied to all databases, except IEEE Xplore because it turned up fewer results. Consequently, it assesses 274 scientific papers before the critical review of the final 66 relevant papers. We publicly provide artifacts of the review for validation and possible reproducibility. The findings show that new NLP tasks are emerging in the battery domain, which facilitate materials discovery and other stages of the life cycle. Notwithstanding, challenges remain, such as the lack of standard benchmarks. The proposed TLP framework, which incorporates agentic AI and optimized prompts, may be apt for tackling some of the challenges.
We present Agaram 2.5B, a Tamil-first language model trained from scratch that converses in Tamil script, romanized Tanglish and English, and calls tools. Our first model, Agaram 1.5B, guided the new base, tokenizer and fine-tuning recipe: we pretrained a new 2.53B-parameter base on 50 billion tokens, 58% of them Tamil, with a tokenizer that needs 1.55 tokens per Tamil word where Qwen2.5's needs 9.44. A four-day public preview of our first 2.5B chat model showed where it fell short with real users, and we rebuilt supervised fine-tuning around what we saw, on 222,662 records. The released model scores above Sarvam-1 in all six translation directions we measured (FLORES chrF 56.9 into Tamil and 55.4 into English), chooses the expected tool on all 25 prompts of a tool set that was used during checkpoint selection and whose schemas appear in training, with exact arguments on 16, and was preferred over the preview model by an LLM judge on 100 sealed human-written prompts (net preference +30 points, 95% interval +14.5 to +45.0). On FLORES it also scores above Google's Gemma 4 E4B-it and our fine-tuned variant of it in both directions, with 2.53B parameters against E4B's 4.5B effective parameters. Factual knowledge and reliable refusal of harmful requests remain open problems, which we set out as future work. The base and chat weights and GGUF files are released under Apache-2.0, and this report gives a full account of the data, training and evaluation as a playbook for building language models for under-represented languages.
In robotic manipulation, the tight coupling between grasping and motion planning often obscures the true source of failure, leading to inefficient trial-and-error. To enable efficient long-horizon manipulation, we propose GTP-FA (Grasp-Then-Plan with Failure Attribution), a task-oriented two-stage 'grasp-then-plan' framework that generates grasp candidates and performs downstream motion planning conditioned on the selected grasp. Given a failed manipulation trajectory, we learn a failure attribution model that generalizes to unseen grasps and produces a stable distribution over failure modes for diagnosis-guided optimization. Based on these attribution results, we then optimize both modules in a diagnosis-driven manner: on the grasping side, we inject task-level priors and risk penalties into grasp candidate scoring and optimization to suppress unstable or task-incompatible grasps; on the planning side, we target high-risk initial states through data collection and fine-tuning to address genuine planning bottlenecks. We evaluate the proposed framework in both simulation and real-robot experiments, and show that GTP-FA improves the corresponding base learners across RL, IL, diffusion-policy, and VLA-based settings, achieving substantially higher overall task success rates.
We measure the emotion geometry of six LLMs spanning Google's Gemma 2 family. Using an extraction pipeline established by Sofroniew et al. (2026) on Claude Sonnet 4.5 and replicated by Hollowell (2026) on Gemma 2 27B base, we compare emotion geometry across base and instruction-tuned variants of Gemma 2 2B, 9B, and 27B. First, all six model cells exhibit the core emotion geometry. A linear map fit between two cells places held-out emotions nearest their own vectors in the target model. Across all fifteen model pairs, fit in one direction each, this failed exactly once, and the median rank was 1 in every pair. Linear maps refit on cleaned embedding-table vectors recover every held-out emotion at rank 1 in all thirty ordered pairs. Second, scale spreads the geometry's variance across more directions without weakening its tie to valence. The leading component's share of variance falls from 31.6 percent (2B) to 29.8 (9B) to 26.7 (27B) in the base models and from 31.9 (2B) to 29.2 (9B) to 27.3 (27B) in the tuned ones, while its correlation with human valence ratings stays between .71 and .78 in all six models. Third, we project each emotion vector onto the output vocabulary and ask at what depth the vector's own emotion word enters the top 10,000 of the 256,000 tokens. In the 2B and 9B models the median emotion word is there from layer 0, while in the 27B models it begins near chance, stays outside the top 10,000 for half the network, and enters at layer 23 in one step. Fourth, tuning leaves every registered analysis-layer measure within its floor. At every size, valence correlation rises and arousal correlation falls by amounts within their floors, the latter against prediction. PC2's variance share also falls, with a pooled decrease significant after correction. Mid-late layer stability exceeds its floor and traces to base 27B's final layer. Code, data, and audit reports are released.
We present GaussianSpline, a dynamic 4D Gaussian Splatting method that models per-Gaussian motion using adaptive, non-uniform piecewise cubic Bézier splines. Unlike approaches based on global deformations, fixed-order polynomials, or uniformly parameterized splines, GaussianSpline jointly learns the Bézier control points and the temporal partition of each Gaussian's trajectory. This adaptive representation allocates greater temporal capacity to highly dynamic regions while maintaining compact trajectories for near-static regions, enabling high-fidelity reconstruction of complex and rapidly changing motion. To jointly achieve globally smooth motion and local adaptivity, we enforce $G^1$ continuity between consecutive Bézier segments using a reflection-based constraint on their control points. This produces smooth, fully differentiable trajectories while preserving the flexibility of non-uniform temporal parameterization. Experiments on the Neural 3D Video benchmark, the Google Immersive dataset, as well as on synthetic scenes demonstrate that GaussianSpline substantially improves reconstruction quality in scenes with highly non-uniform and challenging dynamics, while maintaining competitive performance on general dynamic scenes compared with state-of-the-art 4D Gaussian Splatting methods.
Hang Jiang, Nanxi Liu, Tiantian Wang, Cassandra Overney, Rukun Zhang, Artemisia Luk, Diyi Yang, Deb Roy 等 9 位作者
Given that the Chinese are one of the oldest ethnic groups to immigrate to the US, their cuisine have become central to the US dining scene and an interesting avenue to study how Chinese immigrants have culturally adapted into American life. In that regard, previous research has focused on analyzing the English names that Chinese restaurants adopt, e.g., finding the use of familiar and typical words. To investigate how Chinese immigrants have culturally integrated in the US depending on social, geographic and demographic contexts, we investigate in this work the under-explored relationship between the Chinese and English names of American Chinese restaurants. Specifically, we conduct a study on 3,172 Chinese restaurants (10% of the total on Google Maps) sampled across the US. We present 8-dimension name frames to capture the underlying meanings behind names and create the NameFrames dataset containing 884 annotated parallel Chinese/English names. We also built an NLP pipeline to analyze the semantic discrepancies between names with both frame-based and embedding-based distance metrics. Our frame analysis shows that English names more often emphasize Location and Specialty, while Chinese names more often emphasize Positivity and Ambiance. Furthermore, we observe that for Chinese restaurants located in urban areas, those in counties with a higher concentration of Chinese population, or those that are relatively new, the discrepancies between the English and Chinese names are noticeably larger. Finally, interviews with 10 Chinese restaurant owners reveal additional interesting insights regarding naming conventions.
Search engines increasingly display AI-generated answers above organic links, potentially displacing traffic to upstream publishers. We estimate the impact of Google's AI Overviews (AIO) on Wikipedia's search traffic using AIO's staggered geographic rollout and Wikipedia's multilingual structure. Our difference-in-differences design compares monthly external-search referrals to English Wikipedia articles with referrals to the same articles in German and French, and finds that default AIO availability reduced English search traffic by 5.45% and 4.82%, respectively. Our results suggest that answer-producing digital intermediaries can materially reallocate attention away from informational publishers, with implications for content monetization, search platform design, and policy.
Planetary Geospatial Foundation Models: A New Paradigm for Global Public Health
Arbaaz Muslim, Aviv Slobodkin, Katherine Wheeler-Martin, Eric Zhou, John Brittain, Martin Frasch, Hamsa Subramaniam, Jacob Bien 等 38 位作者
AI 速览：提出并展示地理空间基础模型 PDFM 如何以可泛化的地点表征补充传统流行病学工作流，并在多国多任务案例中改善公共卫生监测。
The efficacy of traditional disease prediction is limited by spatial gaps and temporal lags, which impact the timing and targets of resource deployments. Outbreaks escalate undetected, chronic disease burdens are quantified years later, and at-risk populations in data-sparse regions remain unaddressed. Planetary geospatial foundation models complement existing epidemiological workflows to provide operational improvements, encoding multimodal search, mobility, and environmental signals into generalizable place representations. As illustrations of this complementarity, we present independent global health case studies of Google Earth AI's Population Dynamics Foundation Model (PDFM) -- a foundation model for geospatial inference -- across four domains (vaccine-preventable, communicable, noncommunicable, maternal mental health), five tasks (spatial extrapolation, interpolation/nowcasting, probabilistic forecasting, prospective forecasting, risk stratification), and four countries (USA, Canada, Mexico, and the Democratic Republic of the Congo). Across these case studies, PDFM addresses critical surveillance gaps across domains: improving US-Canada border MMR vaccination coverage predictions by capturing cross-border behavioral spillovers domestic models miss; nowcasting cardiovascular disease to accelerate data availability; enhancing short-term municipal Mexican dengue forecasts for timely outbreak vector control; improving forecasts of cholera hotspots; and adding a transferable signal to individual-level postpartum-depression risk prediction in US states the model had never seen, while not replacing individual socioeconomic data or closing demographic screening gaps. Together, these results showcase capabilities of geospatial foundation models for public health surveillance.