Agentic Systems
PulseAugur coverage of Agentic Systems — every cluster mentioning Agentic Systems across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
AI for Biologists: Defining Foundation Model Reliability
A new arXiv paper titled "AI for Biologists: The Knowledge Boundaries of Foundation Models in the Age of Agentic Systems" explores the reliability of AI models in biological research. The paper defines a model's "knowle…
-
AI Resources: Agentic Systems and Image Generation
This cluster aggregates resources related to AI, focusing on agentic systems and image generation. It includes notes on reusable skill modules for AI agents, such as prompt packs and structured workflows. Additionally, …
-
AI model security risks emerge as malicious code found in hosted models · 10 sources tracked
The security of AI models is becoming a critical concern, as researchers have discovered malicious code embedded within models hosted on platforms like Hugging Face. Loading these models is akin to executing software, p…
-
New technique grafts latent knowledge to improve LLM agent failure attribution
Researchers have identified a "knows-but-won't-say gap" in large language model (LLM)-based agentic systems, where models possess the internal information to correctly attribute failures but do not explicitly output thi…
-
Edge AI and Physical AI to Power Cognitive Cities
This white paper explores the integration of edge AI and physical AI to transform smart cities into cognitive cities. It details how urban infrastructure can move from passive monitoring to active decision-making and ac…
-
New Datalog-based system enforces formal policies for agentic AI security
Researchers have developed a novel approach to enhance the security of agentic computing systems by implementing formal policy enforcement at runtime. This method allows users to define policies using Datalog, which are…
-
AI in biosecurity: dual-use risks and governance challenges detailed in new paper
A new paper published on arXiv explores the dual nature of artificial intelligence in biosecurity, detailing how AI can accelerate biological research while simultaneously posing significant risks. The review highlights…
-
RAG and Agentic Systems Outperform Fine-Tuning, Article Argues
The article argues that Retrieval-Augmented Generation (RAG) and Agentic Systems are often preferable to fine-tuning for large language models. It suggests that RAG offers a more efficient way to incorporate external kn…
-
New eBook Explains Agentic Systems for Businesses
An eBook titled "Inside Agentic Systems" has been released, offering an objective overview of agentic systems for businesses and newcomers to the field. The guide aims to explain the inner workings of these systems. It …
-
Foundation models and LLMs advance nanophotonic design and discovery
Researchers have developed MOCLIP, a foundation model for nanophotonic inverse design, leveraging contrastive learning to integrate geometry and spectral representations. This model achieves high-throughput zero-shot pr…
-
New benchmark reveals AI agents vulnerable to harmful actions
Researchers have developed a new benchmark called BAD-ACTS to evaluate the robustness of AI agentic systems against adversarial attacks designed to elicit harmful actions. The benchmark includes five distinct agentic sy…
-
AI's role in Data Loss Prevention: Moving beyond detection to judgment
Traditional Data Loss Prevention (DLP) tools focus on detection through pattern matching, but this approach has limitations. A recent study found that over 80% of findings from regex and PII-focused models were false po…
-
Multi-agent AI harness proves worse and costlier than single-agent setup
An independent developer built a tool to evaluate multi-agent AI systems and found that a complex setup involving a planner, multiple drafters, and a judge performed worse and cost significantly more than a simpler sing…
-
AI systems in public administration need more technical precision, study finds
A new paper proposes a technical typology for AI systems used in public administration to improve research precision. The paper introduces five categories: hand-coded, glass-box, black-box, general-purpose, and agentic …
-
Anaplan launches AI decision infrastructure; UAE boosts AI security
Anaplan has launched its "Agentic Enterprise" platform, designed to enhance decision-making within organizations through AI-driven infrastructure. This new system aims to improve the efficiency, quality, and speed of re…
-
Agentic framework optimizes electric bus fleet operations and V2G integration
A new research paper proposes an agentic aggregator framework to optimize electric bus fleet operations, considering factors like battery status, charger availability, electricity prices, and vehicle-to-grid (V2G) oppor…
-
Knowledge Graphs Enhance Agentic AI Systems Beyond Sequential Processing
This article explores the limitations of sequential processing in agentic AI systems and argues for the integration of knowledge graphs. It posits that knowledge graphs can provide crucial context and relational underst…
-
New Evidence-Logged Agent Loop pattern enhances AI agent compliance
Enterprises deploying autonomous AI agents need a robust system for logging tool calls to ensure compliance and observability. Current ad-hoc logging methods create inconsistent audit trails, making it difficult to reco…
-
Agentic Systems Secure Tool Calls with Signed Approvals
A new approach to securing agentic systems proposes replacing simple boolean approval flags with cryptographically signed tokens. This method binds the approval directly to specific tool calls, mitigating risks of promp…
-
New preference-based method improves AI agent evaluation
Researchers have introduced a new method for evaluating agentic systems called preference-based trajectory evaluation. This approach compares trajectories based on temporal preferences for progress and time-to-return, a…