Knowledge catalogue
arXiv:2605.08160v1 Announce Type: cross Abstract: Monitoring archaeological sites at scale is vital for protecting cultural heritage, yet pinpointing when disturbances occur remains difficult because
We integrated FrontierCS into Harbor and are releasing a preview long-horizon agent leaderboard (up to 835 turns, ~200K output tokens) with Kimi K2.6 @Kimi_Moonshot (score 46.9) and Claude Code Opus 4
arXiv:2605.08137v1 Announce Type: cross Abstract: Weight pruning is widely advocated for deploying Large Language Models on resource-constrained IoT and edge devices, yet its impact on model fairness
Parameter Golf brought together 1,000+ participants and 2,000+ submissions to explore AI-assisted machine learning research, coding agents, quantization, and novel model design under strict constraint
arXiv:2605.08599v1 Announce Type: new Abstract: Traditional simulation methods reproduce occurred emergency instances through presetting to assist people in risk assessment and emergency decision-maki
April brings Foundry Local GA for local AI development, GPT-5.5 model support with Tier 5 and Tier 6 default quota in Microsoft Foundry, new tracing paths for Microsoft Agent Framework and hosted agen
arXiv:2601.20164v2 Announce Type: replace-cross Abstract: Prior work suggests that language models, while trained on next token prediction, show implicit planning behavior: they may select the next to
arXiv:2605.09549v1 Announce Type: new Abstract: Adaptive prompting mechanisms have been proposed to enhance vision-language models by dynamically tailoring prompts to inputs. However, in frozen few-sh
arXiv:2605.09109v1 Announce Type: new Abstract: Many continuous-control problems ship with a competent but suboptimal controller (a tuned PID, a hand-designed gait). A growing family of methods uses s
arXiv:2605.08318v1 Announce Type: cross Abstract: We study the problem of architecture selection for deep learning models trained to solve partial differential equations (PDEs), asking when transforme
arXiv:2605.10236v1 Announce Type: cross Abstract: Modern off-policy reinforcement learning algorithms often rely on simple uniform replay sampling and it remains unclear when and why non-uniform repla
When is the last time a general purpose LLM (putting aside hybrid systems like Claude Code with special purpose symbolic harnesses) last completely blew away all competing prior models? GPT-4 relative
arXiv:2605.10176v1 Announce Type: cross Abstract: Natural language interfaces to structured databases are becoming increasingly common, largely due to advances in large language models (LLMs) that ena
arXiv:2605.10171v1 Announce Type: cross Abstract: Scientific peer reviews frequently contain conflicting expert judgments, and the increasing scale of conference submissions makes it challenging for A
arXiv:2605.09860v1 Announce Type: new Abstract: Long-horizon reasoning requires deciding not only what actions to take, but how deeply to commit before the next observation. We formalize this as commi
arXiv:2605.06222v2 Announce Type: replace-cross Abstract: World Action Models (WAMs) have recently emerged as a promising paradigm for robotic manipulation by jointly predicting future visual observat
arXiv:2605.10544v1 Announce Type: new Abstract: Long-context adaptation is often viewed as window scaling, but this misses a token-level supervision mismatch: in packed training with document masking,
arXiv:2605.08563v1 Announce Type: new Abstract: When an LLM agent fails a multi-step tool-augmented task and retries, the failed attempt typically remains in its context window -- contaminating the ne
arXiv:2605.10658v1 Announce Type: new Abstract: Continual learning requires new-task adaptation without damaging previously acquired capabilities. Recent forward-pass and zeroth-order (ZO) results sho
arXiv:2605.10912v1 Announce Type: new Abstract: Large language and vision-language models increasingly power agents that act on a user's behalf through command-line interface (CLI) harnesses. However,
arXiv:2605.09511v1 Announce Type: new Abstract: Many downstream decisions in complex terrain require fast wind estimates at a small number of user-specified locations and heights for a given forecast
This post documents a Wordle game result where the player solved puzzle #1,787 in three attempts, using the color-coded emoji system to show letter placement feedback (black for wrong letters, yellow
This post documents a completed game of Wordle (puzzle #1,788) played by Anthropic, showing the progression of guesses across six attempts with color-coded feedback indicating correct letter positions
arXiv:2605.10434v1 Announce Type: new Abstract: Commercial video generation systems such as Seedance2.0 and Veo3.1 have rapidly improved, strengthening the view that video generators may be evolving i
arXiv:2510.03761v2 Announce Type: replace-cross Abstract: The widespread use of preprint repositories such as arXiv has accelerated the communication of scientific results but also introduced overlook
arXiv:2605.09360v1 Announce Type: cross Abstract: Execution-based evaluation of LLM-generated code implicitly treats successful execution as a proxy for correctness. In scientific simulation, this pro
arXiv:2605.08814v1 Announce Type: new Abstract: Chinese character categories are extremely large, and unseen characters frequently arise in open-world scenarios, making zero-shot Chinese character rec
+1 to this. I was recently on a cross-continental flight without wifi, so I brought up Qwen3.6 & Gemma 4 (via @ollama) in Deep Agents on my laptop. admittedly, they fell over on some more involved/com
arXiv:2605.07066v1 Announce Type: new Abstract: Autonomous systems that build structures from natural-language instructions need reliable spatial reasoning, yet large language models (LLMs) make syste
3 weeks since ml-intern launched and we just hit 1M messages exchanged. that's 3.3 agent-years of ML research in 21 days. 2 months worth of research every day. 17,383 training jobs total. talk about A
arXiv:2602.13837v2 Announce Type: replace Abstract: We study video reconstruction from ultra-low-bitrate representations, where the primary challenge shifts from encoding to decoding. In this regime,
arXiv:2605.06681v1 Announce Type: cross Abstract: A hierarchical ensemble pipeline is introduced to address anomaly detection in multivariate telemetry data provided by European Space Agency (ESA). Th
arXiv:2605.07388v1 Announce Type: new Abstract: Marine debris detection for ocean robot is crucial for ecological protection, yet performance is often degraded by low-quality images with blur, complex
arXiv:2605.07170v1 Announce Type: new Abstract: Metaphor is pervasive in everyday language, yet token-level computational identification of metaphor-related words in Chinese under the MIPVU framework
arXiv:2605.06937v1 Announce Type: new Abstract: This methods article presents a reproducible calibration workflow for prompt-based large language models (LLMs) in structured evidence-synthesis tasks.
arXiv:2601.15507v2 Announce Type: replace Abstract: Recent image generation models produce impressive composites, but often fail to preserve the identity of user-provided content when editing specific
arXiv:2605.06924v1 Announce Type: cross Abstract: Synthesizing consistent and coherent long video remains a fundamental challenge. Existing methods suffer from semantic drift and narrative collapse ov
arXiv:2509.08461v3 Announce Type: replace-cross Abstract: Recent advances in Large Language Models (LLMs) have demonstrated their remarkable capacity to process and reason over structured and unstruct
arXiv:2605.06946v1 Announce Type: cross Abstract: Sequence models face a fundamental tradeoff between memory capacity and computational efficiency. Transformers achieve expressive context modeling at
arXiv:2605.07892v1 Announce Type: new Abstract: Sparse training reduces the memory and computational costs of deep neural networks. However, sparse optimization methods, e.g., those adding an ell_1 pe
Agent view is the best Claude Code native way to manage multiple sessions, kind of like tmux built for CC. We spent a lot of time getting the details right, I hope you enjoy it. New in Claude Code: ag
arXiv:2605.07926v1 Announce Type: new Abstract: As LLM-based agents increasingly rely on external tools, it is important to evaluate their ability to sustain tool-grounded reasoning beyond familiar wo
arXiv:2605.06869v1 Announce Type: new Abstract: AI agent research spans a wide spectrum: from RL agents that learn from scratch to foundation model agents that leverage pre-trained knowledge, yet no u
Agents + file sandboxes are all in the range in 2026 🤖🗃️ This is a nifty reference implementation by @itsclelia showing you how to run your agent over a collection of docs (PDFs, images, Office) with
arXiv:2605.06607v2 Announce Type: replace-cross Abstract: Recent LLM-based agents have closed substantial portions of the scientific discovery loop in software-only machine-learning research, in chemi
Nous Research announced free access to Qwen 3.6 Plus and additional features available through their subscription portal at portal.nousresearch.com. This promotion likely provides users with complimen
arXiv:2602.12162v3 Announce Type: replace Abstract: In structurally constrained molecular optimization, state-of-the-art methods restart an expensive oracle-driven search from scratch for every new in
arXiv:2511.09598v5 Announce Type: replace Abstract: Many real-world applications require solving families of expensive multi-objective optimization problems~(EMOPs) under varying operational condition
@trq212: An Anthropic engineer argues HTML is a better output format for AI agents than Markdown, citing information density, ease of sharing, and two-way interaction — Using Claude Code: The Unreason
arXiv:2605.07125v1 Announce Type: cross Abstract: Sequential recommendation has increasingly shifted toward generative recommenders that combine sequential patterns with semantic item information. Yet
arXiv:2605.07046v1 Announce Type: cross Abstract: Evaluation of large language models (LLMs) is increasingly critical, yet standard benchmarking methods rely on average accuracy, overlooking both the
arXiv:2602.19612v3 Announce Type: replace Abstract: Machine Unlearning (MU) enables Large Language Models (LLMs) to remove unsafe or outdated information. However, existing work assumes that all facts
Across the public sector, the conversation has shifted; we are no longer just talking about the potential of AI, we are already seeing the impact. While visionary leadership and cultural buy-in are cr
arXiv:2605.07937v1 Announce Type: new Abstract: Long-horizon AI agents execute complex workflows spanning hundreds of sequential actions, yet a single wrong assumption early on can cascade into irreve
arXiv:2502.07143v3 Announce Type: replace Abstract: The severe shortage of medical doctors limits access to timely and reliable healthcare, leaving millions underserved. Large language models (LLMs) o
arXiv:2605.07191v1 Announce Type: new Abstract: A recent work shows that Attention Transfer, which transfers only the attention patterns from a pre-trained teacher Vision Transformer (ViT) to a random
arXiv:2605.06834v1 Announce Type: new Abstract: Continual learning research attempts to conserve two fundamental capabilities: new knowledge acquisition and the preservation of previously acquired kno
Daybreak is an automated security system that detects potential security threats, validates their legitimacy, and executes response actions with minimal human intervention. Developed by OpenAI, it rep
arXiv:2605.07706v1 Announce Type: new Abstract: Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning of large models by decomposing weight updates into low-rank matrices, significantly r
arXiv:2605.07021v1 Announce Type: new Abstract: Reasoning in Large Language Models (LLMs) poses a challenge for oversight as many misaligned behaviors do not surface until reasoning concludes. To addr