Startups1 мин
Astana Hub and Ghent University open a joint laboratory of artificial intelligence and mathematics
Astana Hub and Ghent University open a joint laboratory of artificial intelligence and mathematics Enbekshi QazaQ
Поиск: Astana Hub AI
Publication Catalog
All stories
All published intelligence, sovereign stack research, ecosystem news, and AI foundation model reviews.
2555
Stories
Research1 мин
Google выпустила Harness² — фреймворк для самоулучшения ИИ-агентов
🤖 Google выпустила Harness² — фреймворк для самоулучшения ИИ-агентов Google Research открыла исходный код системы, которая позволяет ИИ-агентам самостоятельно улучшать свою работу без переобучения нейросети. Harness² даёт агентам возможность менять промпты, настройки памяти,…
Telegram — GPT Time | Нейросети | ИИ
Research1 мин
Microsoft acquired Maluuba a year before GPT-1 was released.
Microsoft acquired Maluuba a year before GPT-1 was released. The company's founder admits that he regrets the sale because Maluuba could have become the next OpenAI. Now the same deep learning team is back with a new project called Skyfall AI, and its first model is...
Telegram — Вайб-кодинг
Research1 мин
Anthropic направит часть своих акций в фонд общественной пользы
Anthropic создаёт специальный фонд, который будет получать долю акций компании и увеличивать капитал по мере её роста. Средства планируют направлять на решение социальных проблем, связанных в том числе с развитием искусственного интеллекта и его влиянием на рынок труда, передает…
Tech-news.kz
Research1 мин
Where the Blocks Live: Tier-Aware KV-Cache Routing for vLLM at Scale
Привет, я Бон-хва Пак из Ware. На конференции Pite Wash в Северной Америке в 2026 году я представлю постер под названием «Где живет блок». Когда память GP заполняется, VLAD принудительно переместит кэш KB на CPU или в хранилище. VLAD будет отслеживать местоположение каждого…
YouTube — PyTorch Official
Research1 мин
Joint, Decoupled, or Agent: When Each Multimodal Alignment Paradigm Wins
arXiv:2610.24108v1 Announce Type: new Abstract: Jointly pretrained multimodal encoders are a common way to combine modalities, while alternatives integrate their information during downstream training or inference. When does each approach win? We compare three paradigms,…
arXiv — cs.AI
Research1 мин
Not Every User Says It All: Training Tool-Using Agents Beyond Ideal Simulated Users
arXiv:2610.24294v1 Announce Type: new Abstract: Tool-using agents are trained and evaluated almost exclusively with strong simulated users who state every detail clearly, yet real users are often weaker. We construct weak users by masking sentences of a simulated user's replies,…
arXiv — cs.AI
Research1 мин
Distillation, Self-Training, and the Fragility of Learned Chemical Reasoning
arXiv:2610.24199v1 Announce Type: new Abstract: Automating synthesis planning for small-molecule drug design requires chemical reasoners that are accurate as well as open to inspection and adaptation, yet frontier language models remain unreliable at predicting reaction outcomes…
arXiv — cs.AI
Research1 мин
Self-Improving Video-to-Action Models
arXiv:2610.24177v1 Announce Type: new Abstract: Video generative models are an emerging tool for embodied agents: they can serve as scalable visual planners, generating videos that depict how to complete previously unseen tasks. However, executing these plans typically requires…
arXiv — cs.AI
Research1 мин
GardenBench: A lightweight, daily evaluation of LLM capabilities
arXiv:2610.23866v1 Announce Type: new Abstract: We introduce GardenBench, a benchmark with 11 test suites that evaluate the capabilities of Large Language Models (LLMs) across a range of tasks. By systematically prompting LLMs with various lightweight queries, we document…
arXiv — cs.AI
Research1 мин
Cache2LoRA: Understanding and Improving Context Internalization
arXiv:2610.24107v1 Announce Type: new Abstract: Context-to-LoRA methods encode documents as low-rank parameter updates, allowing language models to answer without retaining the source text in their decoding context. How reliably do these updates preserve facts and support their…
arXiv — cs.AI
Research1 мин
World of SWECraft: Agent Collaboration at Scale via Simulated Software Worlds
arXiv:2610.24061v1 Announce Type: new Abstract: As teams of AI agents take on larger projects and long-standing open problems, agent collaboration offers another dimension for scaling inference-time compute. To understand what makes such collaboration effective and scalable, we…
arXiv — cs.AI
Research1 мин
Persistent Memory and Iterative Improvement in an LLM Agent for Chinese-Character Drawing
arXiv:2610.24090v1 Announce Type: new Abstract: Does persistent memory accumulated from previous attempts help an LLM agent solve new problems, and does memory design affect that benefit? We study programmatic Chinese-character drawing across five agent conditions attempting the…
arXiv — cs.AI
Research1 мин
Who Speaks, Who Authorizes, Who Acts? Understanding Trust Boundaries in Multi-Agent LLMs
arXiv:2610.24298v1 Announce Type: new Abstract: Multi-agent LLM systems exchange messages before acting through tools, but message delivery, signing authority, and tool-side enforcement are separate trust boundaries. We represent them as communication, authorization, and…
arXiv — cs.AI
Research1 мин
Beyond Explicit Intent: Discovering and Shaping Artistic Intent through Human-AI Co-Creation
arXiv:2610.24010v1 Announce Type: new Abstract: Generative AI is becoming an important medium for artistic creation, yet existing interactions often assume that creators can specify intentions in advance and accurately communicate them through text prompts. Artistic intentions,…
arXiv — cs.AI
Research1 мин
Cortex: A Fixed-Point Theory of Governed Coding Agents - Technical Report
arXiv:2610.24255v1 Announce Type: new Abstract: A coding agent combines a large language model (LLM) with a harness that plans, edits, and executes code. We study Cortex, a supervisory layer for pre-execution governance, requirement tracking, and iterative validation and repair.…
arXiv — cs.AI
Research1 мин
Can Compact Open-Weight Large Language Models Recover Mobility Events from Incomplete Trajectories? A Controlled Evaluation Using Air Trips
arXiv:2610.24266v1 Announce Type: new Abstract: Incomplete mobility trajectories can obscure the observations needed to identify important travel events and their endpoints. This study evaluates whether compact, locally deployed large language models (LLMs) can recover missing…
arXiv — cs.AI
Research1 мин
From SEC Filings to Financial Signals: Evaluating LLMs on Stock Market Reactions and Analyst Forecasts
arXiv:2610.24268v1 Announce Type: new Abstract: Form 10-Q MD&A sections combine long-form management narrative with accounting and forward-looking information. Benchmarks for financial filings commonly evaluate question answering or numerical reasoning, leaving unclear whether…
arXiv — cs.AI
Research1 мин
TorchTPU - A Hybrid Eager/Compiled Architecture for Native PyTorch Execution on TPU
arXiv:2610.24161v1 Announce Type: new Abstract: 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…
arXiv — cs.AI
Research1 мин
Agents' Last Doubt: Knowing When to Act and When to Step Back in Tool-Using Agents
arXiv:2610.24194v1 Announce Type: new Abstract: Tool-using LLM agents repeatedly face decision points under uncertainty, choosing between autonomous action, additional information gathering, and human escalation. These decisions require balancing task success against interaction…
arXiv — cs.AI
Research1 мин
What Social Media Teaches Us About Accumulative AI Risks
arXiv:2610.24006v1 Announce Type: new Abstract: Prominent AI safety discussions emphasize existential threats and catastrophic events, such as AI-enabled nuclear conflict or engineered biological outbreaks. This position paper argues that there is another type of AI risk that…
arXiv — cs.AI
Research1 мин
AM-Bench: A Unified Taxonomy and Evaluation Suite for Agentic Misalignment
arXiv:2610.24264v1 Announce Type: new Abstract: As LLM-based agents are deployed with increasing autonomy in real-world environments, a critical failure mode emerges: agentic misalignment, where agents choose harmful actions over failure. However, existing safety benchmarks…
arXiv — cs.AI
Research1 мин
Training-Free Prompt Selection for Efficient Reinforcement Learning With Partial Rollouts
arXiv:2610.24229v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has emerged as a prominent paradigm for training reasoning models, but its computational cost makes efficient data selection increasingly important. Since training examples…
arXiv — cs.AI
Research1 мин
Unified Multiparadigm Fusion Intelligence for Rare Disease Diagnosis
arXiv:2610.24024v1 Announce Type: new Abstract: Rare diseases (RDs) affect an estimated 300 million people globally, leaving patients to navigate a protracted diagnostic odyssey constrained by clinical scarcity and fragmented medical knowledge. Artificial intelligence (AI) has…
arXiv — cs.AI
Research1 мин
LogicVal:A Logic-Formalized Benchmark for Constraint Satisfaction in Text-to-Image Generation
arXiv:2610.24051v1 Announce Type: new Abstract: Text-to-Image (T2I) generative models have made significant progress in producing high-fidelity images from natural-language prompts. Alongside this progress, numerous benchmarks have been developed to evaluate capabilities such as…
arXiv — cs.AI
Research1 мин
A Conditional No-Hair Theorem for Transformers: How Language Models Forget Their Prompt
arXiv:2610.24133v1 Announce Type: new Abstract: We ask how much an autoregressive transformer retains of its prompt after generating $t$ tokens. Consider two different prompts followed by the same continuation text, so that any difference in their hidden states comes from the…
arXiv — cs.AI
Research1 мин
SteerSearch: Interactive Search Reasoning with Mid-Response Steering
arXiv:2610.24287v1 Announce Type: new Abstract: Search-augmented language models increasingly serve as long-horizon research agents, yet most open training methods assume a fixed user goal throughout search. We study mid-response steering, where users redirect a model during…
arXiv — cs.AI
Research1 мин
Probing Ambiguity: Adversarial Intent Resolution for Robust Text-to-SQL
arXiv:2610.24248v1 Announce Type: new Abstract: Text-to-SQL systems typically assume a one-to-one mapping between natural-language queries and correct SQL translations. In practice, however, many queries are inherently ambiguous, admitting multiple valid interpretations that…
arXiv — cs.AI
Research1 мин
A Controlled LLM Workflow for Evidence-Grounded Diagnosis of Risk Score Differences
arXiv:2610.23962v1 Announce Type: new Abstract: We propose a controlled Large Language Model (LLM) workflow for explaining risk score differences using similar historical transactions. The workflow combines reference retrieval and contrastive SHapley Additive exPlanations (SHAP)…
arXiv — cs.AI
Research1 мин
Doing Less with More: A First-Principles Exploration of the Suitability of Agentic Computing over Alternative Architectural Choices
arXiv:2610.23921v1 Announce Type: new Abstract: There is growing interest in automating business activities with Agentic Artificial Intelligence (AI) due to latter's seeming ease of use. Never has it been easier, or costlier, to do less with more. However, little is known about…
arXiv — cs.AI