Scientific and engineering progress is fundamentally a long-horizon iterative process: proposing changes, running experiments, measuring outcomes, and continuously refining artifacts. Yet existing benchmarks for frontier models primarily evaluate either single-turn responses or short-horizon agent trajectories, failing to capture the challenges of sustained iterative improvement over extended time horizons.
When an AI agent calls an API and hits a validation error, it needs more than what went wrong -- it needs what to do next. A self-reflective API returns, on validation failure, a machine-readable recovery\_feedback.
Generative visual models fundamentally struggle with precise spatial control. This arises from a core disconnect: models can process textual descriptions of space but cannot directly map numerical coordinates onto the 2D image canvas.
Agent Skills today consist largely of free-form prose requiring the agent to read, interpret, and re-derive how to act in every session. This imposes two compounding costs: reduced reliability on implementation-heavy tasks, and difficulty in skill creation and improvement, since editing prose is a fragile process that both humans and agents struggle with, particularly for domain-specific procedural knowledge underrepresented in model training.
How to set the rules that keep agents effective and out of trouble The post What AI Agents Should Never Do on Their Own appeared first on Towards Data Science .
By Sara Nobrega
We present Archi, an open-source, end-to-end framework for scientific collaborations that combines the systematic ingestion and organization of heterogeneous data sources with the deployment of configurable, private, and extensible agents that retrieve and reason over them. An instance of Archi has been deployed for the Computing Operations team of the CMS experiment at CERN's LHC since February 2026 as a support agent for technical operators, offering retrieval and analysis capabilities by combining documentation, historical data, and live monitoring systems.
Temporal credit assignment is central to both biological and artificial intelligence, yet its interaction with non-linear function approximation is poorly understood. We identify a systematic failure mode in deep reinforcement learning (RL) termed Trace-Mediated Peak Bias (TMPB).
As agents grow more capable, legal-domain LLM agents promise to turn document-heavy matters into reviewable work products -- yet reliable deployment faces three obstacles: no large-scale evidence on how today's strongest model-and-harness combinations behave on end-to-end legal matters; no agent architecture adapted to the legal vertical, only general-purpose harnesses; and, in a setting that keeps shifting with new facts, authorities, and deadlines, no mechanism for systems to learn from their own outcomes. We address each.
Purpose The WHO's COVID-19 non-pharmaceutical interventions (e. g.
Large Language Models (LLMs) have become increasingly adopted in daily applications, with deep research standing out as a particularly important capability. Unlike traditional question-answering (QA) tasks, deep research report generation lacks definitive ground-truth, making reward design inherently unverifiable and limiting effective reinforcement learning.
We present AgentJet, a distributed swarm training framework for large language model (LLM) agent reinforcement learning. Unlike centralized frameworks that tightly couple agent rollouts with model optimization, AgentJet adopts a decoupled multi-node architecture in which swarm server nodes host trainable models and run optimization on GPU clusters, whereas swarm client nodes execute arbitrary agents on arbitrary devices.
arXiv:2606. 02914v1 Announce Type: new Abstract: Background: Oral diseases affect nearly 3.
By Sema Helali, Lina Abu Nadab, Sausan Alqawas, Alaa Abd-Alrazaq, Faleh Tamimi, Rafat Damseh
arXiv:2603. 03480v2 Announce Type: replace Abstract: We study reinforcement learning with delayed state observation, where the agent observes the current state after some random number of time steps.
By Harin Lee, Kevin Jamieson
arXiv:2602. 12430v4 Announce Type: replace-cross Abstract: The transition from monolithic language models to modular, skill-equipped agents marks a defining shift in how large language models (LLMs) are deployed in practice.
By Renjun Xu, Yang Yan
arXiv:2606. 02979v1 Announce Type: cross Abstract: We present a novel compact deep multi-task learning model to handle various autonomous driving perception tasks in one forward pass.
By Oskar Natan, Jun Miura
arXiv:2606. 03963v1 Announce Type: cross Abstract: Deep reinforcement learning has shown strong potential for enabling autonomous robots to learn complex navigational tasks.
By Roohan Ahmed Khan, Yasheerah Yaqoot, Muhammad Ahsan Mustafa, Dzmitry Tsetserukou
arXiv:2605. 27762v2 Announce Type: replace Abstract: We present PEAM, a Parametric Embodied Agent Memory framework in Minecraft that transforms agent memory from inference-time retrieval into parameter-resident skills internalized through experience.
By Yuchen Guo, Junli Gong, Weicheng Wang, Hongmin Cai, Yiu-ming Cheung, Weifeng Su
arXiv:2606. 03609v1 Announce Type: cross Abstract: Embodied agents that navigate cities rely on world models that predict how their surroundings will change as they move.
By Xuhui Lin, Stephen Law, Nanjiang Chen, Kunyao Li, Tao Yang
arXiv:2606. 03980v1 Announce Type: new Abstract: Reward models (RMs) provide critical feedback signals for LLM post-training, notably in reinforced fine-tuning (RFT) and reinforcement learning (RL) pipelines.
By Tao Chen, Gangwei Jiang, Pengyu Cheng, Siyuan Huang, Yihao Liu, Jingwei Ni, Jiaqi Guo, Mengyu Zhou, Kai Tang, Junling Liu, Qinliang Su, Xiaoxi Jiang, Guanjun Jiang
arXiv:2606. 02644v1 Announce Type: cross Abstract: Agentic scaffolds have dramatically improved LLM performance on complex, long-horizon tasks, yielding both broad benefits and amplified risks in domains like cybersecurity.
By Eliot Krzysztof Jones, Mateusz Dziemian, Matt Fredrikson, J Zico Kolter