AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchTowards Effective Orchestration of AI x DB Workloads
arXiv:2603.03772v2 Announce Type: replace-cross Abstract: AI-driven analytics are increasingly crucial to data-centric decision-making. Executing relational and AI operators in separate runtimes prevents the database opt…
- arXiv cs.AIresearchSeparators in Enhancing Autoregressive Pretraining for Vision Mamba
arXiv:2603.03806v2 Announce Type: replace-cross Abstract: The state space model Mamba has recently emerged as a promising paradigm in computer vision, attracting considerable attention for its efficient handling of long-…
- arXiv cs.AIresearchSeeing the imagined: latent functional alignment in visual imagery decoding from fMRI data
arXiv:2604.15374v2 Announce Type: replace-cross Abstract: Recent progress in visual brain decoding from fMRI has been enabled by large-scale datasets such as the Natural Scenes Dataset (NSD) and powerful diffusion-based…
- arXiv cs.AIresearchFaithful Autoformalization via Roundtrip Verification and Repair
arXiv:2604.25031v3 Announce Type: replace-cross Abstract: When an LLM formalizes natural language, how do we know the output is faithful? We propose a roundtrip verification approach which does not require ground-truth a…
- arXiv cs.AIresearchLearning Dynamic Evidence Routes for Vision Transformer Probing
arXiv:2605.00915v3 Announce Type: replace-cross Abstract: Probing frozen vision transformers typically uses permutation-invariant aggregation (GAP or $\texttt{[CLS]}$), treating patch tokens as an unstructured set. Conte…
- arXiv cs.AIresearchLLM Ghostbusters: Surgical Package Hallucination Suppression via Adaptive Unlearning
arXiv:2605.01047v2 Announce Type: replace-cross Abstract: Hallucinations remain an unsolved problem for LLMs, and package hallucinations are a particularly dangerous instance of this phenomenon. Package hallucinations oc…
- arXiv cs.AIresearcheXplaining to Learn (eX2L): Regularization Using Contrastive Visual Explanation Pairs for Distribution Shifts
arXiv:2605.06368v2 Announce Type: replace-cross Abstract: Despite extensive research into mitigating distribution shifts, many existing algorithms yield inconsistent performance, often failing to outperform baseline Empi…
- arXiv cs.AIresearchUnbiased Gradients, Moving Stability Boundaries: Exact Mini-Batch Geometry in Linear Self-Attention
arXiv:2605.21292v2 Announce Type: replace-cross Abstract: Unbiased stochastic gradients can match the full gradient in expectation while changing finite-step stability. We study this effect in one-layer linear self-atten…
- arXiv cs.AIresearchLearning Urban Access Costs from Origin-Destination Flows via Inverse Optimal Transport
arXiv:2606.14157v2 Announce Type: replace-cross Abstract: Cities deliver basic services through mixed public-private facility networks, including schools, clinics, transit providers, and subsidized service points. In the…
- arXiv cs.AIresearchWhen Good Verifiers Go Bad: Silent Negative Transfer in Verifier-Guided VLM Training
arXiv:2606.14629v3 Announce Type: replace-cross Abstract: Verifier reliability is not portable across tasks. A verifier-guided self-DPO pipeline with genuine held-out gains on MathVista (+9.6 points on self-training data…
- arXiv cs.AIresearchFinRED: An Expert-Guided Benchmark Generation and Evaluation Framework for Financial LLM Red-Teaming
arXiv:2606.19887v4 Announce Type: replace-cross Abstract: Existing safety benchmarks target general adversarial scenarios but miss finance-specific risks. Financial LLMs face regulatory compliance violations, fraud facil…
- arXiv cs.AIresearchExplanation-Guided Medical Named Entity Recognition with Stability and Boundary Awareness for Atopic Dermatitis
arXiv:2606.22886v2 Announce Type: replace-cross Abstract: Objective: This study aims to improve the reliability and robustness of medical named entity recognition (NER) in Chinese atopic dermatitis (AD) clinical texts th…
- arXiv cs.AIresearchSOLAR: AI-Powered Speed-of-Light Performance Analysis
arXiv:2606.26383v2 Announce Type: replace-cross Abstract: How fast could a deep-learning model run on target hardware, and how far is today's implementation from that limit? These questions are central to software, hardw…
- arXiv cs.AIresearchConditional Co-Ablation: Recovering Self-Repair Backups in Transformer Circuits
arXiv:2607.01940v3 Announce Type: replace-cross Abstract: Mechanistic interpretability seeks to explain transformer behavior through circuits: sets of internal components that causally support a behavior. However, self-r…
- arXiv cs.AIresearchReasonLab: A Controlled and Auditable Evaluation of Prompting Techniques for Multiple-Choice QA
arXiv:2607.14109v2 Announce Type: replace-cross Abstract: Probing the capabilities of Large Language Models (LLMs) and building robust solutions for Multiple-Choice Question Answering (MCQA) remain central challenges in…
- arXiv cs.AIresearchFrom Plausible to Actionable: A Position on LLM Self-Explanations
arXiv:2607.15957v4 Announce Type: replace-cross Abstract: Large Language Models (LLMs) can generate natural language explanations that rationalize their own decisions, a phenomenon commonly referred to as self-explanatio…
- arXiv cs.AIresearchA Methodology for Auditable Trustworthiness Levels in AI Lifecycle Governance
arXiv:2607.16130v2 Announce Type: replace-cross Abstract: AI governance increasingly requires judgments about whether an AI system remains adequately trustworthy over time, whether observed changes are tolerable, and how…
- arXiv cs.AIresearchSemi-Automated Detection of Gaps in LLM Security Knowledge
arXiv:2607.18496v4 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used for a range of software, hardware and human-centered security tasks. Consequently, LLM performance on security…
- arXiv cs.AIresearchParameter-Free Dynamic Regret under Heavy-Tailed Noise
arXiv:2607.27073v3 Announce Type: replace-cross Abstract: We study online convex optimization with one unbiased stochastic subgradient per round and noise having a finite $p$-th central moment, where $p\in(1,2]$ is unkno…
- arXiv cs.AIresearchCorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation
arXiv:2608.03079v2 Announce Type: replace-cross Abstract: Breast core needle biopsy (CNB) is central to breast cancer diagnosis yet remains challenging because limited tissue sampling, lesion heterogeneity, and subtle mo…
- arXiv cs.AIresearchAre Concept Bottleneck Models Effective as Decision-Support Systems?
arXiv:2608.25581v2 Announce Type: replace-cross Abstract: Concept Bottleneck Models (CBMs) are interpretable-by-design neural networks that detect human-understandable concepts from the input and use them to generate pre…
- arXiv cs.AIresearchQuery-Side Attacks on GNN-Based KGQA: Tracing Failures from Entity Linking to Answer Generation
arXiv:2608.25922v2 Announce Type: replace-cross Abstract: GNN-based Knowledge Graph Question Answering (KGQA) pipelines process queries through four discrete stages: entity linking, subgraph retrieval, GNN reasoning, and…
- arXiv cs.AIresearchGVS5H: Zero-Shot Self-Orchestration with Ledger-Based Control for Improved LLM Coding Performance
arXiv:2608.26480v2 Announce Type: replace-cross Abstract: Frontier coding performance is typically attained with large, costly proprietary models. We introduce ledger-based zero-shot self-orchestration (GVS5H), a trainin…
- arXiv cs.AIresearchCompositional Failure in Audio-Visual LLMs: Late-Layer Prior Dominance Under Cross-modal Conflict
arXiv:2608.27785v2 Announce Type: replace-cross Abstract: We study audio-visual conflict as a compositional generalization test for AV-LLMs: the model must combine synchronized but semantically incompatible audio and vid…
- arXiv cs.AIresearchIntrinsic Interaction Geometry Controls the Low-Rank Complexity of Softmax Attention
arXiv:2608.28150v2 Announce Type: replace-cross Abstract: How much matrix rank is required to preserve every bounded value output of normalized softmax attention? We study the unrestricted maximum-row-\(\ell_1\) approxim…
- arXiv cs.AIresearchHBQ: Hierarchical Scaling Block Quantization with Hardware-Efficiency-Aware Design for Accurate LLM Inference
arXiv:2609.00450v2 Announce Type: replace-cross Abstract: Block Quantization (BQ) enables efficient LLM inference by quantizing both weights and activations, but its design space remains underexplored. Through hardware-a…
- arXiv cs.AIresearchTree species mapping in Denmark: A comparison of spectral-temporal features with geospatial foundation model embeddings
arXiv:2609.03480v3 Announce Type: replace-cross Abstract: We map tree species across Denmark using National Forest Inventory plots and EO data, while evaluating the potential of foundation models for large-scale forest c…
- arXiv cs.AIresearchMonadic Second-Order Logic in HOL: Deep and Shallow with Automated Faithfulness (Extended Preprint)
arXiv:2609.07345v3 Announce Type: replace-cross Abstract: In Isabelle/HOL, we apply the deep-and-shallow embedding methodology of our prior work to monadic second-order logic (MSO). Three embeddings are developed side by…
- arXiv cs.AIresearchRice's Theorem under Self-Modification: Elevation Operators and a Normal Form
arXiv:2609.11326v3 Announce Type: replace-cross Abstract: We ask whether it can be certified algorithmically that a self-modifying program keeps a behavioural property, a safety property in the motivating case, after its…
- arXiv cs.AIresearchThe Last AI Built by Humans: Toward Genuine Recursive Self-Improvement
arXiv:2609.11873v3 Announce Type: replace-cross Abstract: Recursive self-improvement (RSI) enables AI systems to turn experience and feedback into persistent changes that improve both their capabilities and the process o…
- arXiv cs.AIresearchChatGPT Images 2.5 in the Wild: A Launch-Period Dataset and Detector Evaluation
arXiv:2609.15100v3 Announce Type: replace-cross Abstract: An image tool can change its underlying generator while retaining its public name, making version attribution from online posts ambiguous. We study this problem a…
- arXiv cs.AIresearchModels as Governed Interfaces for AI-Native MBSE: Read-Side Adequacy and Write-Side Admissibility
arXiv:2609.16252v2 Announce Type: replace-cross Abstract: Machine-readable models such as SysML v2 are now programmatically accessible, and a growing body of work treats that access as the enabling condition for AI parti…
- arXiv cs.AIresearchPhysStream: Streaming Physics-Grounded Video Generation with Structured Scene Memory and Fine-Grained Motion Control
arXiv:2609.17521v2 Announce Type: replace-cross Abstract: Interactive control for video generation is moving from coarse prompts toward fine-grained, physically meaningful manipulation of dynamic scenes. Yet existing con…
- The Verge AIresearchOpenAI wants to consult elite mathematicians about how to not fumble again
After turning a string of spectacular mathematical results into a reputational crisis, OpenAI is consulting human mathematicians to help it figure out a less disastrous path forward. On Monday, the company announced a ne…
- Apple Machine Learning ResearchresearchHow to Guide Your Language Flow
We introduce a new method to guide flow matching models. Our approach, which we call probe guidance, uses the frozen internal states of an existing diffusion model to construct a guidance signal. This works using a simil…
- AI Stack ExchangeresearchWhy is PPO so efficient in early iterations?
I am experimenting with reinforcing learning algorithms, and specifically on-policy learning. Having implemented some variations of known algorithms such as PPO, I am surprised at the efficiency of PPO early on. Sure man…
- AI Stack ExchangeresearchA idea I got from quantum computing about AI , will AGI require partly a quantum computer?
For agi the capability of a agi to take in new experiences, learn from them and use that experience in the future without retraining reminds me of quantum phase kickback.In quantum phase kickback data can enter the contr…
- LessWrongresearchFusion Energy Projects are Not Trying to Imitate the Sun
Post Intro: The quest to use fusion power for break-even electricity generation has sometimes been described as “putting the sun in a jar”. Where by “jar,” we mean a carefully arranged series of coils that produce a powe…
- Latent Spaceresearch🔬 An Oscar, Two Asteroids, and the Algorithm in Your sklearn: John Platt on AI for Science
We talked to Google’s Oscar winning “Giganerd” about automating science, solving climate change, and how future generations can contribute to science in the age of superintelligent AI
- IEEE Spectrum AIresearchWhy Read a Research Paper When You Can Turn It Into an AI Agent?
Have you ever read a paper in Science or Nature and thought, “Man, that research was so cool. I wish I could try that method on my own data,” only to spend a week wrestling with someone else’s undocumented repo, broken d…
- Guardian TechnologyresearchBig tech says AI can find a cure for cancer. So where is it?
Alongside warnings about its ability to destroy humanity, artificial intelligence is also hailed for its potential to revolutionise health treatment. Just how far can it take us? Her name was June. June would be the last…
- The DecoderresearchA tiny software layer from lab-grown neurons promises faster, cheaper AI video
The Biological Computing Co. wants to team up with AWS to sell a text-to-video model that's supposed to run five times faster and 80 percent cheaper thanks to a software layer derived from real nerve cells. The tiny laye…
- LessWrongresearchModern LLMs have tiny GPTs hidden inside them
Experiments into predicting GPT2 completions via Qwen models This is a crosspost from my substack (where I do varied tiny experiments on LLMs and agents). It's also part of Lossfunk , where we're investigating meta-cogni…
- WIRED AIresearchAI Models Built From Rat Brains Just Got Closer to Reality
The Biological Computing Company is bringing its AI tools to Amazon Web Services in a major boost for a once-fringe field that aims to marry nature with code.
- Guardian AIresearchUsing AI to ‘talk to animals’ might make us feel clever – but what, if anything, does it do for them?
As animal behaviorists increasingly use AI in their research, they grapple with how to wield the technology responsibly I’m sitting in a small boat in Shark Bay, Australia, eavesdropping on a conversation between some of…
- The DecoderresearchOpenAI says its internal model solved over 100 long-standing math problems after just a month of training
OpenAI says a new internal model solved more than 100 open math problems after just a month of training. Facing criticism from mathematicians, the company is backing an independent advisory group at the Institute for Adv…
- AI Stack ExchangeresearchIn continuous action spaces, how is the standard deviation, associated with Gaussian distribution from which actions are sampled, represented?
I have a question about implementing policy gradient methods for problems with continuous action spaces. Assume that actions are sampled from a diagonal Gaussian distribution with mean vector $\mu$ and standard deviation…
- AI Stack ExchangeresearchDiffence between AI and brute force in mathematics
I have written two small programs for calculations with Cayley-Dickson Numbers, one with symbols and one with numbers. These programs, written in python, use brute force techniques and do what they are intended to do. Si…
- WIRED AIresearchA New Chatbot Wants to Unlock the Secrets in Tattered Ancient Greek Records
In the hope of uncovering new details about ancient life, researchers have developed a large language model that fills in the gaps in papyrus fragments.
- The DecoderresearchAnthropic is setting up a biology lab where Claude guides robots through drug experiments
Anthropic is building its own biology lab to push AI-driven drug development beyond computer simulations. The article Anthropic is setting up a biology lab where Claude guides robots through drug experiments appeared fir…