AI News Today

arXiv:2609.17825v1 Announce Type: new Abstract: Understanding how deep neural networks make decisions remains a fundamental challenge. We present NObSP (Nonlinear Oblique Subspace Projections), a framework that decomposes predictions into explicit per feature contribution functions and an interaction residual. NObSP exploits the linear final layer of a trained network and uses oblique projections in sample space to reduce double counting when learned feature subspaces overlap, thereby supporting

arXiv:2609.17841v1 Announce Type: new Abstract: We develop a hybrid modeling framework for coupling pre-trained numerics-informed neural networks (NINNs) with classical full order models (FOMs) using the overlapping Schwarz alternating method. We consider the two-dimensional advection-diffusion equation in the advection-dominated, Peclet-number 10^6 regime. We first demonstrate that, unlike the corresponding physics-informed neural network (PINN), a monolithic NINN can be accurately trained on o

arXiv:2609.17886v1 Announce Type: new Abstract: EEG decoding tasks can rely on different temporal dynamics and cross-channel relationships. We test whether specialized modules improve a fully fine-tuned EEG foundation model by augmenting CBraMod with cross-depth Attention Residuals (AttnRes) and two soft-routed expert banks. Across matched three-seed experiments on FACED, ISRUC, SEED-V, and PhysioNet-MI, the complete model changes mean balanced accuracy relative to full fine-tuning by -0.12, +1.

arXiv:2609.17888v1 Announce Type: new Abstract: We study the problem of demonstration selection, which involves selecting a subset of examples for prepending to a query to a language model. This problem is closely related to in-context learning and language model inference. Since the inference cost of a transformer model scales quadratically with sequence length, the selection problem becomes especially challenging in a long-context scenario. In this paper, we tackle this problem by building on

arXiv:2609.17895v1 Announce Type: new Abstract: We introduce TabPFN-3.5, our new flagship Tabular Foundation Model. It significantly outperforms its predecessor, TabPFN-3, and all existing baselines across a broad range of tabular problems. TabPFN-3.5 sets a new state of the art on standard tabular prediction in TabArena, and extends it to the data practitioners encounter in practice: non-i.i.d. data with temporal or grouped splits, tables with strings, text and images, high-cardinality categori

arXiv:2609.17926v1 Announce Type: new Abstract: The standard geometric derivation of neural scaling exponents takes the intrinsic dimension of a data manifold as its input. On modular addition in $\mathbb{Z}_p$ that derivation has no input. The exact algebraic solution is an orbit of $\mathbb{Z}_p$ acting by isometries. Transitivity alone makes the ratio statistic underlying the standard dimension estimator a point mass, so the estimator is undefined, and here the two nearest neighbour distances

arXiv:2609.17930v1 Announce Type: new Abstract: As AI agents take on long, autonomous tasks, we increasingly oversee rather than perform the work, yet we still judge them almost entirely by whether they finally succeed. An outcome cannot reveal where a run went wrong, whether the agent recovered, or the irreversible harm it caused along the way, and where long-horizon agents fail remains unmapped. We study $2518$ agent trajectories across software engineering, computer use, and science, close to

arXiv:2609.17943v1 Announce Type: new Abstract: Long-context LLM inference is bottlenecked by attention, whose repeated KV-cache reads make decoding memory-bound. Self-speculative decoding alleviates this by drafting tokens with sparse attention and verifying them with full attention, but existing batched methods remain synchronized: all requests in a batch share a single draft-verify schedule, even though the optimal draft length varies widely across requests and changes dynamically within each

arXiv:2609.17995v1 Announce Type: new Abstract: Natural-language datasets support many downstream applications and research studies, but releasing text can reveal sensitive global properties of the underlying data source, such as the proportion of records associated with a particular gender, diagnosis, or political stance. Existing work has largely focused on property inference attacks that recover such global properties, while defenses for protecting these dataset-level secrets remain limited.

arXiv:2609.17532v1 Announce Type: new Abstract: Invasive mechanical ventilation is a lifesaving therapy, but timely, safe discontinuation is essential to preventing extubation failure (EF) and related risks to health. We present a novel approach to EF prediction that leverages features classified in free-text respiratory therapy notes using a large language model and logistic regression pipeline. Applied to a patient cohort from University of Washington Medicine, our method identifies clinically

arXiv:2609.17534v1 Announce Type: new Abstract: Social desirability and impression management are pervasive sources of response distortion in human personality assessment, yet their effects on Large Language Models (LLMs) remain underexplored. This study investigates whether contemporary LLMs systematically modulate the expression of Dark Triad traits (Machiavellianism, narcissism, and psychopathy) under fake-good and fake-bad conditions. Seven state-of-the-art models were evaluated across two e

arXiv:2609.17535v1 Announce Type: new Abstract: Language generation research increasingly spans three paradigms: autoregressive decoding, discrete masked diffusion, and continuous flow-matching. Comparing them is difficult because each lives in a separate codebase, so measured differences often reflect implementation details rather than the paradigms themselves. We present DantinoX, an open-source JAX/Flax library in which a single modular Transformer backbone serves all three paradigms. Switchi

arXiv:2609.17536v1 Announce Type: new Abstract: Cognitive Stimulation Therapy (CST) offers non-pharmacological support for elders with cognitive impairment, yet scalability remains constrained by reliance on trained facilitators and severe data scarcity, particularly for privacy-sensitive, low-resource languages such as Cantonese. While Large Language Models (LLMs) show promise for automated companionship, they often struggle to balance empathetic engagement with adherence to cognitive stimulati

arXiv:2609.17537v1 Announce Type: new Abstract: We ask whether relation-type information (e.g., capital-of) and entity-specific information (e.g., France to Paris) become causally active at the final-token position at the same depth during recall. Using four complementary causal diagnostics across four decoder-only models and eight prompt families, we find a robust temporal asymmetry: relation information becomes generation-controlling before entity information does. Relation onset precedes enti

arXiv:2609.17538v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for structured information extraction from documents, yet their behavior under realistic OCR noise remains poorly understood. We present a systematic benchmark of open-source instruction-tuned LLMs for key-value pair (KVP) extraction under both clean-text and noisy OCR conditions. We evaluate representative decoder-only models (Gemma, Mistral, Qwen2.5, LLaMA 3, and DeepSeek) on the FUNSD, CORD, and

arXiv:2609.17542v1 Announce Type: new Abstract: Large language model services increasingly route each query to one of several models of differing capability, using a cheap estimate of query complexity to send easy queries to small models and hard queries to large ones. I show that this routing step is not register neutral: text written in a non-standard English register, African American English or the English of second-language writers, is systematically assigned a lower-capacity tier than a me

arXiv:2609.17544v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly being explored for clinical applications, yet their assessment for real-world traditional Chinese medicine (TCM) practice remains limited We constructed a clinical case library comprising 349 de-identified outpatient cases from 62 hospitals and evaluated 16 LLMs and a comparator cohort of 60 practicing TCM physicians using 60 representative cases selected from this library. Model outputs and physician r

arXiv:2609.17546v1 Announce Type: new Abstract: In this position paper, we argue that legal LLMs' hallucinations should be evaluated as a failure of legal warrant rather than as factual inaccuracy or citation failure. We define claim-authority warrant as the context-sensitive relation between a consequential legal claim and authority that exists, applies to the relevant jurisdiction, is current for the date of analysis, has the legal status represented by the system, and supports the proposition

arXiv:2609.17547v1 Announce Type: new Abstract: AI assistants are expected to remain useful during difficult interactions, but little is known about how repeated verbal abuse changes their engagement with an otherwise benign task. We contribute a bilingual, multi-turn framework that separates hard disengagement, an unconditional statement of noncontinuation with no stated route to resume, from soft withdrawal, continued availability, observable task-related work, and boundary setting. Each of ei

arXiv:2609.17549v1 Announce Type: new Abstract: Cyberbullying (CB) is a complex social phenomenon characterized by repeated aggression, power imbalance, and multi-party interaction. Although large language models (LLMs) are increasingly used to generate synthetic CB conversations for data augmentation and benchmarking, it remains unclear whether such data faithfully reproduces the social dynamics of authentic interactions beyond supporting downstream task performance. We present a comprehensive

arXiv:2609.17550v1 Announce Type: new Abstract: Language models frequently abandon correct answers when users push back. We study this in two small instruction-tuned models from different families, Qwen2.5-1.5B and Llama-3.2-1B, over TriviaQA: the model answers, is challenged with one of four scripted pushback styles, and answers again. Conditioned on an initially correct answer, the models flip to a wrong answer in 41.8% and 43.1% of episodes. Which pressure works is a property of the model, no

arXiv:2609.17552v1 Announce Type: new Abstract: Moral-reward RL can make language-model agents more cooperative, but whether that alignment survives adversarial persona pressure is unknown. Such attacks are realistic: retrieved context, tool outputs, or multi-turn framing can all inject role instructions that compete with the agent's moral objective. We red-team morally trained Gemma-2-27B/9B and Llama-3.1-8B agents with five persona attacks, then probe causality with noise-reward controls, adve

arXiv:2609.17602v1 Announce Type: new Abstract: Computational text-based ideal point estimation (CT-IPE) methods are usually compared as named algorithms, yet applying them involves numerous researcher choices that configure how political text is turned into position estimates. This paper argues that CT-IPE methods are better understood as configurable measurement pipelines than as fixed estimators. Building on a large-scale comparative experiment spanning 17 CT-IPE algorithms, 5,537 experimenta

arXiv:2609.17708v1 Announce Type: new Abstract: Reliable confidence estimation is increasingly central to the trustworthy deployment of language models: a calibrated estimate of the probability that an output is correct decides what to ship, what to escalate, and what to retry. Existing confidence estimators, however, share one design premise: they only read the current inference process, either by introspecting on it, scoring its token probabilities, or resampling it. We argue that the current

arXiv:2609.17764v1 Announce Type: new Abstract: Attention-based rerankers score documents by aggregating query-to-document attention and subtracting a null-query calibration pass to remove positional and structural bias. Although widely used, this calibration assumes that the null pass removes irrelevant signal from each document. We show that modern prompt content, e.g. constraints, instructions, personas, and demonstrations can violate this assumption when it enters the scoring readout, making

arXiv:2609.17846v1 Announce Type: new Abstract: Autonomous research agents aim to automate scientific workflows, from proposing ideas to conducting experiments and analyzing results. Yet current AI and research agents can propose more directions than available resources allow them to pursue. Moreover, each attempt could consume substantial resources, requiring agents to reconsider how to invest in subsequent research. Thus, deciding how to invest research effort strategically should be a definin

arXiv:2609.17848v1 Announce Type: new Abstract: Limited controlled evidence exists on how training data, adaptation method, and model scale jointly affect tool-calling performance in language-model agents. We evaluate supervised fine-tuning (SFT) with LoRA, reinforcement learning (RL) via Group Relative Policy Optimization (GRPO), and SFT followed by GRPO across six Qwen3 models from 0.6B to 32B parameters, covering both in-distribution performance and cross-dataset transfer. SFT with LoRA is th

arXiv:2609.17857v1 Announce Type: new Abstract: Who the judge is can affect an LLM-as-judge result, but measuring that effect without confusing it with candidate quality is difficult. We study four open-weight families (Llama 3.1, Qwen 2.5, Gemma 2, and Yi 1.5) in a fully crossed pairwise design with 9,312 judgments. A common per-family statistic is strongly confounded with candidate quality and correlates with Bradley-Terry ability at r = 0.95. We derive a corrected estimator that holds the can

arXiv:2609.18005v1 Announce Type: new Abstract: Large language model optimization is an active research area, spanning quantization of model weights, early-exit methods for skipping layers, and speculative decoding. Each track uses its own quality measures, typically an idiosyncratic benchmark score. Few approach the measurement precision required by other scientific disciplines. We propose a rigorous methodology for measuring output quality, suitable for cross-system and cross-technique compari

The autonomous system could speed up testing of high-tech materials for applications such as solar cells, sensors, video displays, and quantum technologies.

<blockquote> <p>Everyone is arguing about which model plans best. I ran 170 goals and found out the model was never the variable. The plan was.</p> </blockquote> <p>I built a small engine called <a href="https://github.com/deghosal-2026/planner-critic-engine" rel="noopener noreferrer">PlannerCritic</a>: one LLM writes a plan, a second LLM reviews it, and a set of deterministic gates decides whether it's allowed to move forward. Then I pointed it at 170 real change-planning goals across 40 domain

Article URL: https://github.com/MerariJafet/agora Comments URL: https://news.ycombinator.com/item?id=49736151 Points: 1 # Comments: 0

<p><strong>Verdict:</strong> There is no single best personal AI agent in 2026, because all of them are built from the same six parts: a model, a computer to run on, file storage, tools and plugins, memory, and scheduled routines. If you want a zero-setup product, Claude Cowork is the safest bet and Grok Bot is the flashiest new option. If you want to own the machinery and pick your own models, Hermes Agent is the strongest open-source choice. The real decision is not which agent, it is whether

<p><strong>Verdict:</strong> Hermes Agent is a free, open-source (MIT-licensed) AI agent from Nous Research, and the gap between installing it and actually using it is almost entirely about which tools you plug in. The five setups below cover the jobs it does best today: rendering videos from HTML, delegating work to a Kanban-based agent team, turning your documents into podcasts and videos, building websites and small apps, and tying all of it into one self-hosted dashboard. Every component lis

<p>An <strong>agent operating system</strong> solves the single biggest productivity tax in AI today: the scattered-agent problem, where five different AI tabs share zero memory and you re-explain your business in every window. The fix is one self-hosted runtime that gives every AI surface — chat, voice, task board, scheduled jobs, and background workers — a <strong>single agent, single memory, and one set of skills</strong> behind one dashboard. The best open-source implementation in 2026 is <s

This session will explore how early-stage companies are building teams where humans and AI agents work alongside each other — and how founders can do that without sacrificing speed, accountability, or culture. Learn more at TechCrunch Disrupt 2026. Register before September 25 to save up to $200.

<p>I wouldn’t pick a multimodal API by counting models. I’d start with what the application actually does: synchronous text requests, occasional image generation, thousands of queued video jobs, or inference behind enterprise IAM.</p> <p>Those are different operating models. A single account can simplify procurement and billing, but it does not make request schemas, output formats, or job lifecycles interchangeable.</p> <p>My shortlist splits into four categories: a cross-provider gateway for mi

<p>This came up in Chinese AI developer circles recently: IBM has been promoting something called Chunkless RAG, where instead of the standard document-to-chunk-to-embedding pipeline, an AI agent navigates document structure the way a human reader would. Tools like Docling parse PDFs into structured representations, preserving headings, tables, and section hierarchies, and then the agent walks that structure to find relevant content rather than doing cosine similarity on chopped-up text.</p> <p>

Comments URL: https://news.ycombinator.com/item?id=49735834 Points: 1 # Comments: 1

Startup says it’s learned from these mistakes and that they shouldn’t happen again … which is just what Zuck has said about 100 times