A practical guide to Amazon Nova Multimodal EmbeddingsArtificial Intelligence In this post, you will learn how to configure and use Amazon Nova Multimodal Embeddings for media asset search systems, product discovery experiences, and document retrieval applications.
In this post, you will learn how to configure and use Amazon Nova Multimodal Embeddings for media asset search systems, product discovery experiences, and document retrieval applications. Read More
Microsoft announced today that the Exchange Web Services (EWS) API for Exchange Online will be shut down in April 2027, after nearly 20 years. […] Read More
Someone used AI to build an entire Web platform, which then did something predictable and preventable: It exposed all its data through a publicly accessible API. Read More
Ransomware operators are hosting and delivering malicious payloads at scale by abusing virtual machines (VMs) provisioned by ISPsystem, a legitimate virtual infrastructure management provider. […] Read More
This week didn’t produce one big headline. It produced many small signals — the kind that quietly shape what attacks will look like next. Researchers tracked intrusions that start in ordinary places: developer workflows, remote tools, cloud access, identity paths, and even routine user actions. Nothing looked dramatic on the surface. That’s the point. Entry […]
Scaling In-Context Online Learning Capability of LLMs via Cross-Episode Meta-RLcs.AI updates on arXiv.org arXiv:2602.04089v1 Announce Type: new
Abstract: Large language models (LLMs) achieve strong performance when all task-relevant information is available upfront, as in static prediction and instruction-following problems. However, many real-world decision-making tasks are inherently online: crucial information must be acquired through interaction, feedback is delayed, and effective behavior requires balancing information collection and exploitation over time. While in-context learning enables adaptation without weight updates, existing LLMs often struggle to reliably leverage in-context interaction experience in such settings. In this work, we show that this limitation can be addressed through training. We introduce ORBIT, a multi-task, multi-episode meta-reinforcement learning framework that trains LLMs to learn from interaction in context. After meta-training, a relatively small open-source model (Qwen3-14B) demonstrates substantially improved in-context online learning on entirely unseen environments, matching the performance of GPT-5.2 and outperforming standard RL fine-tuning by a large margin. Scaling experiments further reveal consistent gains with model size, suggesting significant headroom for learn-at-inference-time decision-making agents. Code reproducing the results in the paper can be found at https://github.com/XiaofengLin7/ORBIT.
arXiv:2602.04089v1 Announce Type: new
Abstract: Large language models (LLMs) achieve strong performance when all task-relevant information is available upfront, as in static prediction and instruction-following problems. However, many real-world decision-making tasks are inherently online: crucial information must be acquired through interaction, feedback is delayed, and effective behavior requires balancing information collection and exploitation over time. While in-context learning enables adaptation without weight updates, existing LLMs often struggle to reliably leverage in-context interaction experience in such settings. In this work, we show that this limitation can be addressed through training. We introduce ORBIT, a multi-task, multi-episode meta-reinforcement learning framework that trains LLMs to learn from interaction in context. After meta-training, a relatively small open-source model (Qwen3-14B) demonstrates substantially improved in-context online learning on entirely unseen environments, matching the performance of GPT-5.2 and outperforming standard RL fine-tuning by a large margin. Scaling experiments further reveal consistent gains with model size, suggesting significant headroom for learn-at-inference-time decision-making agents. Code reproducing the results in the paper can be found at https://github.com/XiaofengLin7/ORBIT. Read More
A Novel Framework for Uncertainty-Driven Adaptive Explorationcs.AI updates on arXiv.org arXiv:2509.03219v5 Announce Type: replace
Abstract: Adaptive exploration methods propose ways to learn complex policies via alternating between exploration and exploitation. An important question for such methods is to determine the appropriate moment to switch between exploration and exploitation and vice versa. This is critical in domains that require the learning of long and complex sequences of actions. In this work, we present a generic adaptive exploration framework that employs uncertainty to address this important issue in a principled manner. Our framework includes previous adaptive exploration approaches as special cases. Moreover, we can incorporate in our framework any uncertainty-measuring mechanism of choice, for instance mechanisms used in intrinsic motivation or epistemic uncertainty-based exploration methods. We experimentally demonstrate that our framework gives rise to adaptive exploration strategies that outperform standard ones across several environments.
arXiv:2509.03219v5 Announce Type: replace
Abstract: Adaptive exploration methods propose ways to learn complex policies via alternating between exploration and exploitation. An important question for such methods is to determine the appropriate moment to switch between exploration and exploitation and vice versa. This is critical in domains that require the learning of long and complex sequences of actions. In this work, we present a generic adaptive exploration framework that employs uncertainty to address this important issue in a principled manner. Our framework includes previous adaptive exploration approaches as special cases. Moreover, we can incorporate in our framework any uncertainty-measuring mechanism of choice, for instance mechanisms used in intrinsic motivation or epistemic uncertainty-based exploration methods. We experimentally demonstrate that our framework gives rise to adaptive exploration strategies that outperform standard ones across several environments. Read More
EvoFSM: Controllable Self-Evolution for Deep Research with Finite State Machinescs.AI updates on arXiv.org arXiv:2601.09465v2 Announce Type: replace
Abstract: While LLM-based agents have shown promise for deep research, most existing approaches rely on fixed workflows that struggle to adapt to real-world, open-ended queries. Recent work therefore explores self-evolution by allowing agents to rewrite their own code or prompts to improve problem-solving ability, but unconstrained optimization often triggers instability, hallucinations, and instruction drift. We propose EvoFSM, a structured self-evolving framework that achieves both adaptability and control by evolving an explicit Finite State Machine (FSM) instead of relying on free-form rewriting. EvoFSM decouples the optimization space into macroscopic Flow (state-transition logic) and microscopic Skill (state-specific behaviors), enabling targeted improvements under clear behavioral boundaries. Guided by a critic mechanism, EvoFSM refines the FSM through a small set of constrained operations, and further incorporates a self-evolving memory that distills successful trajectories as reusable priors and failure patterns as constraints for future queries. Extensive evaluations on five multi-hop QA benchmarks demonstrate the effectiveness of EvoFSM. In particular, EvoFSM reaches 58.0% accuracy on the DeepSearch benchmark. Additional results on interactive decision-making tasks further validate its generalization.
arXiv:2601.09465v2 Announce Type: replace
Abstract: While LLM-based agents have shown promise for deep research, most existing approaches rely on fixed workflows that struggle to adapt to real-world, open-ended queries. Recent work therefore explores self-evolution by allowing agents to rewrite their own code or prompts to improve problem-solving ability, but unconstrained optimization often triggers instability, hallucinations, and instruction drift. We propose EvoFSM, a structured self-evolving framework that achieves both adaptability and control by evolving an explicit Finite State Machine (FSM) instead of relying on free-form rewriting. EvoFSM decouples the optimization space into macroscopic Flow (state-transition logic) and microscopic Skill (state-specific behaviors), enabling targeted improvements under clear behavioral boundaries. Guided by a critic mechanism, EvoFSM refines the FSM through a small set of constrained operations, and further incorporates a self-evolving memory that distills successful trajectories as reusable priors and failure patterns as constraints for future queries. Extensive evaluations on five multi-hop QA benchmarks demonstrate the effectiveness of EvoFSM. In particular, EvoFSM reaches 58.0% accuracy on the DeepSearch benchmark. Additional results on interactive decision-making tasks further validate its generalization. Read More
Today’s “AI everywhere” reality is woven into everyday workflows across the enterprise, embedded in SaaS platforms, browsers, copilots, extensions, and a rapidly expanding universe of shadow tools that appear faster than security teams can track. Yet most organizations still rely on legacy controls that operate far away from where AI interactions actually occur. The result […]
The elusive Iranian threat group known as Infy (aka Prince of Persia) has evolved its tactics as part of efforts to hide its tracks, even as it readied new command-and-control (C2) infrastructure coinciding with the end of the widespread internet blackout the regime imposed at the start of the month. “The threat actor stopped maintaining […]
