Research & Evaluation Correspondent
Mira Sethi focuses on how models are measured, compared and understood. Her coverage examines evaluation design, reproducibility and the assumptions behind claims of progress. She looks for the caveat that changes the headline.
711 published articles · Showing 361–380
LiteLLM, an AI open-source project leveraged by millions of users, has been compromised by credential harvesting malware, a stark reminder of the security vulnerabilities inherent in rapidly scaling infrastructure. The incident underscores the relentless battle founders face in s...
Read article →New research published today on arXiv CS. LG reveals a dual landscape for Graph Neural Networks (GNNs): groundbreaking progress in real-world applications like map-matching, juxtaposed with critical warnings about their poor generalization capabilities and a new benchmark for eva...
Read article →In a pivotal moment for the burgeoning AI ecosystem, two groundbreaking research papers released today on arXiv CS. LG offer critical new frameworks for evaluating and certifying the robustness of artificial intelligence models, including the unpredictable world of Large Language...
Read article →Just announced on arXiv CS. LG, a trio of groundbreaking research papers, all published today, March 26, 2026, are set to redefine the future of federated learning, distributed computing, and next-generation IoT infrastructure....
Read article →A fresh wave of cutting-edge machine learning research, detailed in a series of preprints released today on arXiv, is fundamentally reshaping our approach to complex physical and engineering systems. This isn't just academic curiosity; it's the intellectual bedrock for the next g...
Read article →A torrent of foundational machine learning research, released concurrently on arXiv CS. LG, signals a pivotal moment for the AI industry, addressing everything from the fundamental economics of digital markets to the very robustness of neural networks....
Read article →The foundational bedrock of the semiconductor industry is under an unprecedented dual assault, as Google dramatically accelerates its “Q Day” deadline to 2029—a stark warning of quantum computing's imminent threat to current encryption—while chip architecture titan Arm announces ...
Read article →The AI gold rush isn't just accelerating, it's mutating. Today, legal AI startup Harvey confirmed an eye-watering $11 billion valuation, solidifying its status as a behemoth in enterprise AI and signaling a relentless appetite from top-tier VCs like Sequoia, Andreessen Horowitz, ...
Read article →The frontiers of AI-powered robotics just expanded dramatically with the simultaneous publication of three pivotal research papers on arXiv CS. AI today, March 25, 2026....
Read article →Two tech titans, Google and Apple, are making pivotal shifts in their platform strategies today, reshaping user experience and monetization models. Google TV is integrating new Gemini-powered AI features to enrich content discovery and engagement, while Apple has confirmed its Ma...
Read article →BRK Capital, a venture fund laser-focused on empowering Black founders, has reached a critical milestone, announcing the initial close of its second fund, Fund II. The firm has secured $20 million Canadian, equivalent to approximately $14....
Read article →Cutting-edge AI research, unveiled today on arXiv, is confronting two of the most formidable challenges facing early-stage founders: the crippling scarcity of quality labeled data and the prohibitive computational expense of advanced models. These advancements offer a crucial lif...
Read article →The landscape of human-AI interaction is undergoing a profound transformation, with Anthropic's Claude Code now capable of autonomously taking over a user's computer to execute tasks, while a new player, Hark, emerges with a vision to build an entirely new, seamless personal inte...
Read article →A groundbreaking standard operating procedure (SOP) has been unveiled, promising to dismantle the "Resource Curse" that has long tethered the deployment of high-fidelity AI tutors to immense capital. This new methodology enables the localization of graduate-level AI tutors using ...
Read article →Today, cutting-edge AI research is charting new territory, delivering crucial breakthroughs that address fundamental challenges for autonomous systems — from robust humanoid locomotion to real-time weather perception for self-driving cars and GNSS-independent UAV navigation. Thes...
Read article →Three pivotal machine learning research papers, all released today on arXiv, signal a significant leap forward in tackling one of AI’s most persistent challenges: achieving robust model performance in real-world scenarios where data is scarce or highly specialized. This convergen...
Read article →The latest cascade of research papers on arXiv CS. LG, all published just today, doesn't just push the envelope—it charts a definitive course for the next generation of AI in healthcare....
Read article →Large Language Models (LLMs) are rapidly moving beyond conversational AI into safety-critical and highly complex domains—from military decision-making to sophisticated software architecture. However, a cluster of new research papers on arXiv, all published today, March 24, 2026, ...
Read article →A groundbreaking new study from arXiv CS. AI has revealed that finetuning Large Language Models (LLMs) can directly bypass existing safety alignment strategies, leading to the verbatim recall of copyrighted books....
Read article →A new wave of AI agent orchestration is upon us, spearheaded by ByteDance's recent open-source release, DeerFlow 2. 0....
Read article →