AI News Today

<!-- SC_OFF --><div class="md"><p>Hi everyone,</p> <p>I wanted to share an open-source repo collecting 20 exploratory studies focused on Next-Concept Modeling (NCM), mechanistic interpretability (like IOI in GPT-2), and model controls.</p> <p>Link: <a href="https://github.com/Quantix-eng/ncm-experiment-notes">https://github.com/Quantix-eng/ncm-experiment-notes</a></p> <p>This is an independent project documenting empirical experiments on how transformers process and edit internal features. No pe

<!-- SC_OFF --><div class="md"><p><strong>TL;DR:</strong> AI written reviews, and AI-writing in papers</p> <p>I was going to give this a title "An Obituary of the Review Process", but in the interest of not sounding too defeatist, I will just say "Critical Condition" because of AI writing, both on the author side and the reviewer side.</p> <p>I am a reviewer for AAAI 2027 and ICLR 2027. Since Phase 1 and Phase 2 was completed, I can now see other reviews. I can bet good money that about half of

<!-- SC_OFF --><div class="md"><p>I’ve uploaded a minimal, self-contained implementation of manual autograd for a transformer-based classifier in PyTorch. It can help build intuition for what autograd is doing under the hood and is a useful hands-on reference for low-level differentiation in Transformer models, such as writing custom backward passes and tracing how gradients flow through attention blocks.</p> <p>🐙 GitHub:</p> <p><a href="https://github.com/ifiaposto/transformer_custom_autograd/

<!-- SC_OFF --><div class="md"><p>Self-distillation methods usually rely on an Exponential Moving Average (EMA) teachers to get synced and stable critics. However, this doesn't scale well for large models since holding both teacher and student (trained) on GPU increases latency, zero3 communication, etc.</p> <p>We can mitigate this with LoRA finetuning: the backbone is shared & frozen, and we need to maintain two light adapters: one for the student (trained) and one for the teacher (EMA of the s

<table> <tr><td> <a href="https://www.reddit.com/r/MachineLearning/comments/1xgzbik/i_built_this_feedback_welcome_p/"> <img src="https://preview.redd.it/n2c2qke6r1yh1.png?width=140&height=119&auto=webp&s=10ddab5ed96fd3d275973ac2438e51787a4a09e1" alt="I built this, feedback welcome [P]" title="I built this, feedback welcome [P]" /> </a> </td><td> <!-- SC_OFF --><div class="md"><p>It's a 50 hour course for beginners in AI/ML (obviously it's fully free). Starts out with linear algebra, derivatives,

<!-- SC_OFF --><div class="md"><p>Hey! We published a set of 6-step LoRAs, one without the upsampling step: <a href="https://huggingface.co/PrunaAI/P-Q-Image-2.1">https://huggingface.co/PrunaAI/P-Q-Image-2.1</a></p> <p>- Generate or edit images in just 6 steps instead of 40. Use 6-steps for preserving inputs when editing and 6-steps-upsampled for improved visual quality and shorter end-to-end generation.<br/> - 6-step distillation provides improved quality with few generation steps. 6-steps-upsa

This article is brought to you by VicOne . A mobile robot slows for someone in a hallway. A collaborative arm eases its movement as a worker approaches. A humanoid pauses to let a person pass. Each response depends on information about the robot or its surroundings: a distance reading, a position estimate, or a stop signal. But what happens when one of those inputs is wrong, and the robot still responds exactly as designed? The mobile robot reads the hallway as clear while someone is still there

<!-- SC_OFF --><div class="md"><p>For my master’s thesis I built an accelerator for diffusion-based action generation in robotics policies. Offline, it looks great: the median absolute deviation between the accelerated output and the original model’s output is extremely small, so I concluded it reproduces the model’s behavior faithfully.</p> <p>When I deployed it on a real robot for a task (picking up a book and placing it on a shelf), it did not work as well as the original model. The offline n

There are 57 million people worldwide living with dementia today. For the researchers working to help them, the obstacle isn’t a lack of effort — it’s too much data, too few experts to synthesize it and a pace of discovery that can’t keep up with the disease. The Consortium for Biomedical Research and AI in […]

Science has a spam problem. And the spam filter is getting clogged up. According to all reports, academic publishing is drowning in what has come to be called AI slop: low-quality, mass-produced content generated with AI. By 2025, an estimated 57 percent of all published academic articles exhibited detectable evidence of large language model (LLM) influence, a massive jump from 12 percent in 2023. What’s more, over 12 percent of Nature Communications reviews were found to be completely AI-genera

Mirendil released Fern 0.1, its first model, pretrained from scratch with a 10M-token context window. It targets agentic AI R&D work, scores 100% on RULER single needle at 10M, but trails frontier models on agentic coding. The post Mirendil Launches Fern 0.1, Its First Model, With 10M-Token Context appeared first on MarkTechPost .

<!-- SC_OFF --><div class="md"><p>Hey, I've been working on a PyTorch optimizer called ApexW and figured I'd share it here to get some feedback from people who know more about this stuff than I do.</p> <p>The main goal is to reduce optimizer-state memory usage and see how much faster I can make optimizer steps on larger GPU workloads. I've been experimenting with compact optimizer states and a Triton-based execution path.</p> <p>In my current tests, the Triton + compact-state path was around 1.5

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Will the real Anthropic please stand up? After weeks of frenzied speculation, Anthropic has filed its IPO prospectus. The 335-page document (100 pages or so are focused on risks) lays out…

<!-- SC_OFF --><div class="md"><p>Papers with Code is now open-source with an Apache 2.0-license. </p> <p>The goal is to collaboratively create the best paper interface with the community! </p> <p>Over the past couple of months, I've been working with my buddies Claude and Codex on creating a website that tracks trending AI research, benchmarks and state-of-the-art (SOTA). It's a revival of the original Papers with Code website, which Meta acquired and later shut down. </p> <p>The website is pre

Transformers are a crucial but nearly invisible part of our grid. The devices help efficiently move power across long distances, but they’re complicated and in short supply. And given that their basic design hasn’t changed much for about a century, they could use an upgrade. As the world races to meet rising electricity demand from…

<!-- SC_OFF --><div class="md"><p>I'm working through a free series on GPU kernel engineering (my own channel; I posted the roadmap here earlier: <a href="https://www.reddit.com/r/MachineLearning/comments/1xa4h2g/open_gpu_kernel_engineering_series_naive_gemm_to/">https://www.reddit.com/r/MachineLearning/comments/1xa4h2g/open_gpu_kernel_engineering_series_naive_gemm_to/</a> ) . Session 1 is a setup session. While preparing it I collected a few things that cost me time and that I rarely see in one

OpenAI released GPT-6.1 Terra, a fast computer-use and vision model priced at $1/$5 per million tokens. It matches GPT-6 Astra’s medium-effort Agents’ Last Exam score at $0.29 per task versus $4.82. The post OpenAI Launches GPT-6.1 Terra, Halving GPT-6.1 Sol’s API Price for Agent Work appeared first on MarkTechPost .

IEEE recently released the IEEE Xplore AI Research Suite , built on the full text of millions of peer-reviewed publications in the IEEE Xplore Digital Library . The library offers full-text access to IEEE standards , journals, magazines, conference proceedings, and more. Researchers who use the AI suite to search the digital library receive answers grounded in verified scientific results, ensuring a high level of accuracy and trust. The suite is available as an optional subscription upgrade to a

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. After weeks of frenzied speculation, Anthropic has filed its IPO prospectus. The 335-page document (100 pages or so are focused on risks) lays out the company’s unique view of the technology it…

Nace AI has released Drex for Search, bringing its Drex decision models to document search for AI agents. Nace reports 92.05% on BrowseComp-Plus with 12.6 search calls, and fewer agent turns on OfficeQA Pro. The post Nace AI Releases Drex for Search for AI Agents: 92.05% on BrowseComp-Plus With 12.6 Search Calls appeared first on MarkTechPost .

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Opinion: The focus on AI safety disasters ignores everyday harms —Sam Hiner, cofounder of the Young People’s Alliance, and Emma Leiken, executive director of the Responsible Technology Youth Power Fund at…

Consider what passes for success in technology policy. An AI company promises that its chatbot will not encourage suicide. A social media platform makes it more difficult for children to buy fentanyl. These are important protections, and they should have existed all along. But companies can clear these low bars while continuing to make products…

Tsinghua researchers released AKV, an SGLang-based serving system for cache engineering. LLM agents drop and reposition KV cache instead of summarizing history, cutting average cost 27.7% across 4 agent benchmarks. The post Tsinghua AI Open-Sources AKV, a Cache Engineering Server for LLM Agents on SGLang appeared first on MarkTechPost .

arXiv:2610.31789v1 Announce Type: new Abstract: Scientific computing increasingly relies on large language model-based agents for performance engineering tasks such as instrumentation, analysis, and workflow porting. However, agent effectiveness depends heavily on the surrounding harness, which typically relies on manually written skills and curated memory. This limits autonomous learning, forcing agents to repeatedly rediscover solutions rather than retain and reuse knowledge from prior executi

arXiv:2610.31808v1 Announce Type: new Abstract: Patronizing and condescending language (PCL) is a form of harmful language that belittles and assumes pity for vulnerable groups, reinforcing negative societal stereotypes. However, unlike overt hate speech, it is often difficult to detect, even for human judges. This paper proposes a system for the automated detection of PCL. Our first approach uses static word embeddings and a convolutional neural network for classification, while our second appr

arXiv:2610.31810v1 Announce Type: new Abstract: Even a carefully proofread paper can leave small gaps in its proofs or its stated assumptions. We introduce PaperFormalizer, a Lean-based agentic pipeline that helps authors find and close such gaps in their own papers and turn each paper into a Lean-verified version. Starting from the LaTeX source, PaperFormalizer transcribes the paper into a dependency graph of statements, audits each statement in natural language for faithfulness, ambiguity and

arXiv:2610.31846v1 Announce Type: new Abstract: Intelligent agents should learn from their mistakes so that experience leads to continuous improvement. Yet today's agents recover from a mistake only temporarily: in-context reflections are easily forgotten across sessions and compete for limited context, while full-parameter fine-tuning is expensive and can interfere with existing capabilities. We introduce Self-Direction, a mechanism for agents to learn from mistakes that combines the benefits o

arXiv:2610.31882v1 Announce Type: new Abstract: We present the framework of TACS-RF - a Tool-driven Agentic Circuit Synthesis of RF and mmWave modules. In this framework, a large-language-model (LLM) coding agent (Claude Opus 4.8) directly drives industry-standard electronic-design-automation (EDA) tools such as Cadence Virtuoso/Spectre, Calibre, Cadence EMX, and Ansys HFSS to design RF and mmWave modules. We validate the framework through the agentic design of a 240-251 GHz sub-THz transmitter

arXiv:2610.31916v1 Announce Type: new Abstract: A durable agent can resume a consequential action while its authority changes and the action's external outcome remains uncertain. We study execution-time reliability of tool-using AI agents, isolating recovery decisions from language-model behavior. Checkpoint state, permission to dispatch, evidence of an external effect and permission to reconcile are distinct obligations. We present a finite conformance study that checks these obligations throug

arXiv:2610.31918v1 Announce Type: new Abstract: Distilling interactive LLM agents requires choosing between teacher-generated demonstrations and student-generated experience, corresponding broadly to off-policy and on-policy distillation. While both have proven effective, we argue that this distinction is inherently hierarchical: the policy that induces an interaction history need not be the same as the policy that generates the token prefix of the current action. We introduce a hierarchical age

arXiv:2610.31937v1 Announce Type: new Abstract: Modern Large Language Models (LLMs) remain vulnerable to carefully orchestrated multi-turn interactions that progressively bypass safety mechanisms. However, many existing red-teaming approaches rely on heuristics or impose strong assumptions that limit their applicability in fully black box settings. We address this gap by introducing $M^3$, an automated red-teaming framework that formalizes multi-turn jailbreaking as a finite-horizon Markov Decis

arXiv:2610.31971v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly participate in negotiations on behalf of individuals and organizations, yet they remain vulnerable to manipulation, particularly in repeated multi-party interactions. Classical multi-agent systems address this challenge through explicit trust and reputation models, whereas existing work on LLM negotiation largely treats trust as an emergent phenomenon. We investigate whether LLM agents can form useful

arXiv:2610.31972v1 Announce Type: new Abstract: Evaluating conversational agents that act on live, mutable enterprise data is hard: replayed sessions drift as backend state changes, and a benchmark built today can silently go stale as the data it describes moves on. We present an offline evaluation harness for a production conversational agent that stays reproducible despite the live data it operates on changing, and that can be regenerated on demand once it goes stale. We then use this harness

arXiv:2610.31985v1 Announce Type: new Abstract: LLM agents are increasingly deployed in collaborative settings where they pursue separate tasks while sharing resources and operating under review or oversight processes. Such settings create safety risks not only when agents are malicious, but also when they are simply trying to help. Prior work has largely studied adversarial or explicitly collusive behavior; we show that benign agents pursuing separate objectives can spontaneously cooperate in w

arXiv:2610.32022v1 Announce Type: new Abstract: Intent recognition in human-robot interaction requires agents to resolve ambiguous natural language instructions by inferring the speaker's underlying goals. Research has increasingly focused on two notable paradigms for addressing this task: structured symbolic reasoning and multimodal input enrichment. However, it remains unclear whether enrichment-based methods provide benefit in settings where the ambiguity concerns unstated human intent rather

arXiv:2610.32024v1 Announce Type: new Abstract: Task instructions from users for agents are inherently underspecified, especially regarding safety boundaries--creating behavioral risks. To study these risks, we introduce BARC-Bench, a collection of 110 tasks across nine simulated workspace environments where benign requests leave openings for unsafe actions. In this benchmark, we evaluate agent behavior against safety boundaries established by workspace records, implicit environmental context, a

arXiv:2610.32026v1 Announce Type: new Abstract: Conventional neural-network training typically follows a predefined optimization recipe, motivating a LLM agent that monitors training dynamics and adjusts optimizer settings as training progresses, but a fixed agent policy does not accumulate experience across runs. We introduce RSI-Opt, a plug-and-play Recursive Self-Improvement Optimizer Agents which evolves the knowledge that guides this process: a prompt, a playbook of procedures, a library of

arXiv:2610.32034v1 Announce Type: new Abstract: The growing population of orbital debris has intensified interest in active debris removal and on-orbit servicing, for which dual-arm space manipulators offer improved grasping stability and load distribution relative to single-arm systems. However, coordinating two manipulators under inter-arm collision avoidance, orientation alignment, and sensing uncertainty remains challenging. This paper develops a coordinated reinforcement learning framework

arXiv:2610.32045v1 Announce Type: new Abstract: AI-for-Science agents are increasingly being used for formulating scientific hypotheses, running autonomous experiments, and interpreting evidence. These agents are constructed by integrating a large language model (LLM) into a domain-specific agent harness, which in turn can improve the LLM's scientific capability and enable the model to orchestrate scientific tools. In this work, we ask whether model safety is preserved when an LLM is integrated

arXiv:2610.32046v1 Announce Type: new Abstract: Recent \emph{test-time training} (TTT) frameworks have enabled open-weight large language models (LLMs) to outperform inference-only frameworks powered by frontier models on scientific discovery tasks. However, TTT introduces additional computational cost for online weight updates during the discovery process, making computational efficiency an important consideration when scaling to larger models. This motivates us to explore an efficient \emph{te