Category
2,215 papers
Large language models have improved chiefly by scaling one monolithic computation. We argue the binding constraint is no longer missing components—language, reasoning, and knowledge are all present—but a missing composition in which they form a closed loop. We propose the Core Ring, an architecture with no single core....
AbstractUMOSU/UMOKWAI — Foundational Phenomena and Domains of Complexity: What an AI Needs to Answer Reliably**This deposit collects the first systematic, phenomenon-by-phenomenon analysis conducted within the UMOSU/UMOKWAI ontological and epistemic-operational model. The model classifies reality through six foundation...
Large language models (LLMs) are often memory-bandwidth bound during autoregressive decoding, so reducing weight storage does not automatically produce parallel speedup when the compressed representation is irregular. We present SSQT, a post-training framework that jointly applies hardware-aligned structured sparsifica...
Large language models have improved chiefly by scaling one monolithic computation. We argue the binding constraint is no longer missing components—language, reasoning, and knowledge are all present—but a missing composition in which they form a closed loop. We propose the Core Ring, an architecture with no single core....
Reach audiences
Advertise in front of researchers, engineers, and readers.
The performance of single-GPU LLM inference is characterized by strong cross-layer interactions spanning model architecture, runtime scheduling, operator execution, and GPU microarchitecture. Unfortunately, a unified understanding of single-GPU LLM inference bottlenecks is still lacking due to two limitations: the lack...
AbstractUMOSU/UMOKWAI — Foundational Phenomena and Domains of Complexity: What an AI Needs to Answer Reliably**This deposit collects the first systematic, phenomenon-by-phenomenon analysis conducted within the UMOSU/UMOKWAI ontological and epistemic-operational model. The model classifies reality through six foundation...
As Large Language Models (LLMs) are increasingly used as a primary source of information and advice, understanding their alignment to humans in terms of values becomes a pressing concern. A growing literature has leveraged large scale surveys to investigate to what extent LLMs'and humans'stated values and opinions alig...
The state of the art in AI for requirements engineering research leading up to the transformation, before reviewing advances in LLMs, and two subsequent research areas: prompt programming and generalist SE agents, which combine multiple LLM advances to yield semi-autonomous processes that complete SE tasks.
The need for custom validation capabilities to ensure that models produced by both humans and LLMs are trustworthy enough to serve as the foundation of the MBSE ecosystem is addressed.
OntoSTL defines a bidirectional transformation between STL and OWL and applies equivalence-preserving canonicalization rules, including implication elimination, supported Negation Normal Form (NNF) conversion, Boolean simplification, associative flattening, and deterministic operand ordering.
Business Process Modeling (BPM) is a key and complex activity in Software and Systems Engineering. As a consequence, any relevant assistance integrated into the BPM tooling can be highly beneficial. Historically, Natural Language Processing (NLP) techniques have already been used in this context. However, the recent de...
Effective model-driven engineering of software-intensive systems relies on appropriately designed conceptual modeling languages. A well-designed modeling language includes a set of concepts that users of the language find to be relevant for describing phenomena of interest in a domain, as well as a vocabulary of lexica...
Learn how Google DeepMind and GovTech Singapore uses lightweight on-device Gemma models to intercept conversational scams in private chats with zero cloud data egress.
StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners. It models how individuals may engage with tutors, giving feedback to improve tutor performance across subjects like chess, writing, and mathematics. The post Simulating student learning for better AI tutoring appeared first on Microsoft Research.
By rethinking how large cloud computing systems operate, Associate Professor Christina Delimitrou seeks to make data centers more energy efficient.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
