Researchers baked 3x inference speedups directly into LLM weights — without speculative decoding

Researchers from the University of Maryland, Lawrence Livermore, Columbia and TogetherAI have developed a training technique that triples LLM inference speed without auxiliary models or infrastructure ...
Communications of the ACM

Formal Reasoning Meets LLMs: Toward AI for Mathematics and Verification

A marriage of formal methods and LLMs seeks to harness the strengths of both.
Global Times

StepFun speech model tops global Speech Reasoning with 96.4 percent accuracy

Chinese large language model startup StepFun's speech model Step-Audio R1.1 (Realtime) ranked first globally in the Speech Reasoning category with an accuracy rate of 96.4 percent, according to data ...
EurekAlert!

MathEval: a comprehensive benchmark for evaluating large language models on mathematical reasoning capabilities

This study introduces MathEval, a comprehensive benchmarking framework designed to systematically evaluate the mathematical reasoning capabilities of large language models (LLMs). Addressing key ...
SiliconANGLE

Harmonic AI raises $120M at $1.45B valuation to advance mathematical reasoning

Artificial intelligence for formal mathematical reasoning startup Harmonic AI Inc. announced today that it has raised $120 million in new funding on a $1.45 billion valuation. The funding is intended ...
Business Wire

Harmonic Builds Momentum Towards Mathematical Superintelligence with $120 Million Series C

Ribbit Capital Leads Round at $1.45B Valuation of Math-Based AI Venture; Emerson Collective Joins Existing Backers Including Sequoia & Kleiner Perkins PALO ALTO, Calif.--(BUSINESS WIRE)--Harmonic, the ...
TechCrunch

Harmonic, the Robinhood CEO’s AI math startup, launches an AI chatbot app

Harmonic, an AI startup co-founded by Robinhood CEO Vlad Tenev, announced Monday the beta launch of an iOS and Android chatbot app where users can access its AI model, Aristotle. With this launch, the ...
VentureBeat

New AI architecture delivers 100x faster reasoning than LLMs with just 1,000 training examples

Singapore-based AI startup Sapient Intelligence has developed a new AI architecture that can match, and in some cases vastly outperform, large language models (LLMs) on complex reasoning tasks, all ...
IEEE

Thinking on Context: Inductive Relation Prediction Guided by the Reasoning Ability of Large Language Models

Abstract: Inductive relation prediction aims to predict missing connections between entities unseen during training. Recent approaches adopt binary (positive or negative) training labels, which ...
Scientific Research Publishing

Zalaghi, H. and Khazaei, M. (2016) The Role of Deductive and Inductive Reasoning in Accounting Research and Standard Setting. Asian Journal of Finance & Accounting, 8, 23-37.

ABSTRACT: The paper explores how integrating alternative fuels and renewable energy technologies—like solar, wind, and geothermal—into the UK’s sustainable design can promote sustainable design in the ...
Scientific American

At Secret Math Meeting, Researchers Struggle to Outsmart AI

In a weekend in the spring of 2025, a clandestine mathematical conclave convened. Thirty of the world’s most renowned mathematicians traveled to Berkeley, Calif., with some coming from as far away as ...