AI News Today

arXiv:2607.01586v1 Announce Type: cross Abstract: Vision-language-action models (VLAs) have recently advanced robotic manipulation, yet the effects of different robot-data pre-training paradigms remain difficult to compare because existing models often differ in architecture, data, action space, and evaluation protocol. We present VLAFlow (Vision-Language-Action Flow), a unified flow-matching framework for controlled comparison of VLA training objectives. Using a heterogeneous robot corpus, OXEM

arXiv:2607.01627v1 Announce Type: cross Abstract: Accurate protein-protein interaction (PPI) prediction is central to functional genomics, disease mechanism discovery, and drug development. A difficult setting arises when candidate interactions include proteins that have no observed PPI edges during training, where models relying on network topology alone often lose useful context. This paper presents \method, a multimodal representation framework for cold-start PPI prediction. \method\ combines

arXiv:2607.01689v1 Announce Type: cross Abstract: Model merging aims to combine existing single-task solutions into a multi-task solution without additional data-driven fine-tuning.~Most existing approaches achieve this using geometric properties of local solution spaces. However, such geometric views provide limited guidance for scoring how statistically useful each task-specific update direction is across tasks during merging. To address this, we formulate model merging from a new perspective

arXiv:2607.01702v1 Announce Type: cross Abstract: Recently, speech classification methods have gained widespread adoption in intelligent gadgets. Current study indicates that backdoor attacks provide a substantial security concern to these models, underscoring the pressing necessity to investigate additional potential attack techniques to expose and prevent such risks. This work discusses the vulnerability of current speech triggers to detection by deep neural network defenders and introduces th

arXiv:2607.01736v1 Announce Type: cross Abstract: We study how to predict the downstream closed-loop performance of a learned latent world model from validation-time diagnostics alone. Choosing the right checkpoint from a world-model training run is difficult: validation loss and multi-step prediction RMSE keep improving long after closed-loop performance has collapsed. We present a suite of structural validation-time diagnostics drawn from optimal-control theory and apply them to Gymnasium's Lu

arXiv:2607.01741v1 Announce Type: cross Abstract: Reinforcement Learning (RL) is a sequential decision-making framework in which an agent learns optimal policies through interaction with an environment by maximizing cumulative rewards. Among RL methods, Bayesian Reinforcement Learning (BRL) addresses common practical challenges related to data scarcity by leveraging prior knowledge about the environment and sequential belief updates. However, most BRL approaches require an explicit likelihood fu

arXiv:2607.01751v1 Announce Type: cross Abstract: Existing medical video benchmarks primarily evaluate whether a model produces the correct answer, but rarely assess whether it answers at the right time. In real clinical settings, AI systems must decide not only what to predict, but also when to answer, defer judgment, or proactively raise alerts. This creates a critical gap between benchmark evaluation and deployment requirements. We present MedStreamBench, a benchmark for time-aware medical vi

arXiv:2602.05999v4 Announce Type: replace Abstract: How does the amount of compute available to a reinforcement learning (RL) policy affect its learning? Can policies using a fixed amount of parameters, still benefit from additional compute? The standard RL framework does not provide a language to answer these questions formally. Empirically, deep RL policies are often parameterized as neural networks with static architectures, conflating the amount of compute and the number of parameters. In th

arXiv:2602.10431v4 Announce Type: replace Abstract: Large language models (LLMs) demand substantial computational and memory resources, posing challenges for efficient deployment. Two complementary approaches have emerged to address these issues: token-adaptive layer execution, which reduces floating-point operations (FLOPs) by selectively bypassing layers, and quantization, which lowers memory footprint by reducing weight precision. However, naively integrating these techniques leads to additio

arXiv:2603.12676v3 Announce Type: replace Abstract: Generalizing neural surrogate models across different PDE parameters remains difficult because changes in PDE coefficients often make learning harder and optimization less stable. The problem becomes even more severe when the model must also predict beyond the training time range. Existing methods usually cannot handle parameter generalization and temporal extrapolation at the same time. Standard parameterized models treat time as just another

arXiv:2603.27631v2 Announce Type: replace Abstract: Self-supervised pre-training, where large corpora of unlabeled data are used to learn representations for downstream fine-tuning, has become a cornerstone of modern machine learning. While a growing body of theoretical work has begun to analyze this paradigm, existing bounds leave open the question of how sharp the current rates are, and whether they accurately capture the complex interaction between pre-training and fine-tuning. In this paper,

arXiv:2604.21254v3 Announce Type: replace Abstract: LLM architecture research generally aims to maximize model quality subject to fixed compute/latency budgets. However, many applications of interest such as edge and on-device deployment are further constrained by the model's memory footprint, thus motivating parameter-efficient architectures for language modeling. This paper describes a simple architecture that improves the parameter-efficiency of LLMs. Our architecture makes use of looped Tran

arXiv:2605.22472v2 Announce Type: replace Abstract: Winner-take-all (WTA) networks constitute a central circuit motif in cortical networks of the brain. In addition, WTA-like activations are abundant in modern deep learning models in the form of the softmax activation for example in attention layers of transformers. While their role in the extraction of latent factors has been studied for relatively simple generative models, their role in the context of highly non-linearly entangled latent facto

arXiv:2606.00342v2 Announce Type: replace Abstract: We study the problem of differentially private (DP) $k$-means clustering in Euclidean space. Previous solutions rely on summing the private data directly, which induces a sensitivity proportional to the domain. We introduce PE-means, an extension of the private evolution (PE) algorithm (an increasingly popular method for synthetic data generation), to the problem of $k$-means clustering. The key advantage of PE is that it only computes a privat

arXiv:2503.07570v3 Announce Type: replace-cross Abstract: Deep learning, when integrated with a large amount of training data, has the potential to outperform machine learning in terms of high accuracy. Recently, privacy-preserving deep learning has drawn significant attention of the research community. Different privacy notions in deep learning include privacy of data provided by data-owners and privacy of parameters and/or hyperparameters of the underlying neural network. Federated learning is

arXiv:2504.03711v2 Announce Type: replace-cross Abstract: Artificial intelligence (AI)-driven electronic design automation (EDA) techniques have been extensively explored for VLSI circuit design applications. Most recently, foundation AI models for circuits have emerged as a new technology trend. Unlike traditional task-specific AI solutions, these new AI models are developed through two stages: 1) self-supervised pre-training on a large amount of unlabeled data to learn intrinsic circuit proper

arXiv:2510.16031v2 Announce Type: replace-cross Abstract: Machine learning-based precipitation nowcasting relies on high-fidelity radar reflectivity sequences to model the short-term evolution of convective storms. However, the development of models capable of predicting extreme weather has been constrained by the coarse resolution (1-2 km) of existing public radar datasets, such as SEVIR, HKO-7, and GridRad-Severe, which smooth the fine-scale structures essential for accurate forecasting. To ad

arXiv:2512.07541v3 Announce Type: replace-cross Abstract: Inspired by graph-based methodologies, we introduce a novel graph-spanning algorithm designed to identify changes in both offline and online data across low to high dimensions. This versatile approach is applicable to Euclidean and graph-structured data with unknown distributions, while maintaining control over error probabilities. Theoretically, we demonstrate that the algorithm achieves high detection power when the magnitude of the cha

arXiv:2512.10485v2 Announce Type: replace-cross Abstract: Vulnerability detection methods based on deep learning (DL) have shown strong performance on benchmark datasets, yet their real-world effectiveness remains underexplored. Recent work suggests that both graph neural network (GNN)-based and transformer-based models, including large language models (LLMs), yield promising results when evaluated on curated benchmark datasets. These datasets are typically characterized by consistent data distr

arXiv:2601.01095v3 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) have achieved impressive progress in vision-language reasoning, yet their ability to understand temporally unfolding narratives in videos remains underexplored. True narrative understanding requires grounding who is doing what, when, and where, maintaining coherent entity representations across dynamic visual and temporal contexts. We introduce NarrativeTrack, the first benchmark to evaluate narrat

arXiv:2602.23561v2 Announce Type: replace-cross Abstract: Symbolic regression (SR) has gained recent traction in AI-driven scientific discovery for learning closed-form physical laws. Yet existing methods are dominated by heuristic search or data-intensive approaches that often assume low-noise regimes and lack principled uncertainty quantification, while fully probabilistic SR formulations remain scarce. We introduce a scalable probabilistic framework for SR, VaSST, based on variational inferen

arXiv:2603.09046v3 Announce Type: replace-cross Abstract: Device-side Large Language Models (LLMs) have witnessed explosive growth, offering higher privacy and availability compared to cloud-side LLMs. During LLM inference, both model weights and user data are valuable, and attackers may even compromise the OS kernel to steal them. ARM TrustZone is the de facto hardware-based isolation technology on mobile devices, used to protect sensitive applications from a compromised OS. However, protecting

arXiv:2603.24016v2 Announce Type: replace-cross Abstract: Multi-Object Tracking (MOT) has traditionally focused on a few specific categories, restricting its applicability to real-world scenarios involving diverse objects. Open-Vocabulary Multi-Object Tracking (OVMOT) addresses this by enabling tracking of arbitrary categories, including novel objects unseen during training. However, current progress is constrained by two challenges: the lack of continuously annotated video data for training, an

There's a scientific reason. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

I see sandboxes are getting cool and common. But I wanted to do so for free and with a local focus. I've been a one person operation lately so cloud based repos seemed a bit overkill and the surface is small enough that it was a fun experiment. If you have any thoughts let me know :) Comments URL: https://news.ycombinator.com/item?id=48770421 Points: 2 # Comments: 0

Imagent gives AI agents the ability to generate images, video, and speech as a first-class step in their workflows, behind a single interface that hides the differences between providers and models Comments URL: https://news.ycombinator.com/item?id=48770383 Points: 3 # Comments: 0

Working with a customer to implement RevTurbine (revturbine.com). We are implementing a reverse trial which is triggered by the key onboarding action (connecting the user's trading account), i.e. used as an incentive to complete it. The trial is on a separate non-public tier (subset of lowest paid plan features) which provides great flexibility: play with limits to control AI costs, segment however you like (higher limits for stronger prospects), etc - which is all in line with the customer prom

Here are my biggest gripes with AI coding assistants right now: Obsessed with reinventing the wheel. You'll often find it writing three duplicate functions for the exact same feature in a single file. Why? Because it's terrified of blowing up the context window, so it only reads a fraction of a large file and completely misses the existing functionality. Why are files so bloated in the first place? Because AI prefers adding new code over modifying existing code, and it rarely deletes anything. A

Article URL: https://praxiscurrents.substack.com/p/moneyball-for-physical-ai Comments URL: https://news.ycombinator.com/item?id=48770157 Points: 2 # Comments: 0

Article URL: https://github.com/lotus-data/lotus Comments URL: https://news.ycombinator.com/item?id=48770058 Points: 3 # Comments: 0

<p>Powered by chipmakers Samsung Electronics and SK Hynix, South Korea is seeing a surge in wealth, but there are questions over who gets to share in the profits</p><p>When South Korea’s most high-profile divorce case returned to court last month, the lawyers were arguing not just about the breakdown of a relationship, but also the exact date at which to value shares in one specific company.</p><p>The judges’ decision in Seoul could change the value of business tycoon Chey Tae-won’s assets by bi

Workers who use OG crowdsourcing platform say AWS is closing accounts

Then what? What do you plan to do? The strategy is the "Everyone becomes Levels-io". The takeaway is "marketing games is big now, the only thing." 2026 was always giong to be my Year of Marketing, as BrowserBox reached stability last year and it's bascially: small edges, maintenance, minor updates, customer deployments and customer tweaks. But I have so many product ideas. I am just cranking them out. If you check my GH activity, even before AI I was always "cranking them out" - just, less capab

<p>Follow the day’s latest updates</p><ul><li><p>Get our <a href="https://www.theguardian.com/email-newsletters?CMP=cvau_sfl">breaking news email</a>, <a href="https://app.adjust.com/w4u7jx3">free app</a> or <a href="https://www.theguardian.com/australia-news/series/full-story?CMP=cvau_sfl">daily news podcast</a></p></li></ul><p>A teenager has been charged with murder after a 15-year-old boy was discovered with fatal stab wounds outside a community medical centre.</p><p>AAP reports the boy was f

<p>I use a lot of little web tools. A JSON formatter here, a cron parser there, a "cm to inches" when I am reading a spec written by someone on the other side of the planet. And every single time, the same friction: the first result is buried in ads, a cookie wall pops up, and half of them want me to sign in just to format some text.</p> <p>So I built the thing I actually wanted: <a href="https://yourhack.ai" rel="noopener noreferrer">yourhack.ai</a>, a directory of 300+ free web tools. No ads.

Article URL: https://www.youtube.com/watch?v=JXJf7vL8k94 Comments URL: https://news.ycombinator.com/item?id=48769755 Points: 2 # Comments: 1

<p>The mistake with coding agents is treating them like a single chat window.</p> <p>You paste a task. The agent writes a patch. You check it. Something is missing. You paste another task. Then another. Then CI fails. Then review comments come in. Then you realize you are still the project manager, the QA loop, the scheduler, and the person remembering what every thread was supposed to do.</p> <p>The better workflow is to make one Codex thread responsible for orchestration.</p> <p>Not just the c

Article URL: https://www.bbc.com/news/articles/cj6gr0xkyr3o Comments URL: https://news.ycombinator.com/item?id=48769696 Points: 23 # Comments: 14

I tried to code a systems language called Tig https://github.com/alonsovm44/tc-lang It is cool and it taught me a lot of CS and programing topics, but i ran out of tokens and since i have no monery for cascade code, the project stagnated. I was ashamed of myself so i turned to good old man programs, tutorials in youtube and i tried writing a lexer for a calculator in raw C. And so far i've been successful. Moral of the story: Dont depend on AI, learn to code by yourself. It will take longer, it