Generative AI

Report says Minnesota workers face highest generative AI exposure in the Midwest

A report from North Star Policy Action says Minnesota workers have the highest generative AI exposure in the Midwest and the 10th-highest...

Navigating Recent Developments in Generative AI and Trade Secret Protection

[2601.03127] Unified Thinker: A General Reasoning Modular Core for Image Generation

Abstract page for arXiv paper 2601.03127: Unified Thinker: A General Reasoning Modular Core for Image Generation

[2602.15914] Steering Dynamical Regimes of Diffusion Models by Breaking Detailed Balance

This paper explores how breaking detailed balance in generative diffusion processes can enhance reverse processes while maintaining stati...

[2602.15894] Quality-constrained Entropy Maximization Policy Optimization for LLM Diversity

This paper presents Quality-constrained Entropy Maximization Policy Optimization (QEMPO), a method to enhance diversity in large language...

[2602.15874] P-RAG: Prompt-Enhanced Parametric RAG with LoRA and Selective CoT for Biomedical and Multi-Hop QA

The paper introduces P-RAG, a novel hybrid architecture that enhances Retrieval-Augmented Generation (RAG) for biomedical question answer...

[2602.15895] Understand Then Memory: A Cognitive Gist-Driven RAG Framework with Global Semantic Diffusion

The paper presents CogitoRAG, a novel Retrieval-Augmented Generation framework that enhances semantic integrity and reasoning in language...

[2602.15889] Evidence for Daily and Weekly Periodic Variability in GPT-4o Performance

This article investigates the temporal variability in the performance of the GPT-4o model, revealing significant daily and weekly pattern...

[2602.16570] Steering diffusion models with quadratic rewards: a fine-grained analysis

This article presents a detailed analysis of sampling from reward-tilted diffusion models, focusing on quadratic rewards and their comput...

[2602.16548] RIDER: 3D RNA Inverse Design with Reinforcement Learning-Guided Diffusion

RIDER introduces a novel framework for 3D RNA inverse design using reinforcement learning, significantly enhancing structural similarity ...

[2602.15867] Playing With AI: How Do State-Of-The-Art Large Language Models Perform in the 1977 Text-Based Adventure Game Zork?

This paper evaluates the performance of state-of-the-art Large Language Models (LLMs) in the 1977 text-based adventure game Zork, reveali...

[2602.15863] Not the Example, but the Process: How Self-Generated Examples Enhance LLM Reasoning

This paper explores how self-generated examples can enhance reasoning in Large Language Models (LLMs), emphasizing the process of example...

[2602.15862] Enhancing Action and Ingredient Modeling for Semantically Grounded Recipe Generation

This paper presents a novel framework for improving recipe generation from food images by enhancing action and ingredient modeling, addre...

[2602.15861] CAST: Achieving Stable LLM-based Text Analysis for Data Analytics

The paper presents CAST, a framework designed to improve the stability of LLM-based text analysis in data analytics by enhancing output c...

[2602.16498] Fast and Scalable Analytical Diffusion

The paper presents GoldDiff, a novel framework for analytical diffusion that enhances scalability and speed in generative modeling by dyn...

[2602.15856] Rethinking Soft Compression in Retrieval-Augmented Generation: A Query-Conditioned Selector Perspective

The paper presents SeleCom, a novel selector-based soft compression framework for Retrieval-Augmented Generation (RAG), addressing limita...

[2602.16473] Synthesis and Verification of Transformer Programs

This paper presents C-RASP, a programming language for transformers, and introduces new algorithms for its synthesis and verification, en...

[2602.15854] Decoupling Strategy and Execution in Task-Focused Dialogue via Goal-Oriented Preference Optimization

This paper presents Goal-Oriented Preference Optimization (GOPO), a new framework for enhancing task-oriented dialogue systems by decoupl...

[2602.15853] A Lightweight Explainable Guardrail for Prompt Safety

The paper presents a Lightweight Explainable Guardrail (LEG) method for classifying unsafe prompts in AI systems, utilizing a multi-task ...

[2602.16449] GICDM: Mitigating Hubness for Reliable Distance-Based Generative Model Evaluation

The paper presents GICDM, a method to mitigate hubness in distance-based evaluations of generative models, enhancing reliability and alig...

[2602.15851] Narrative Theory-Driven LLM Methods for Automatic Story Generation and Understanding: A Survey

This survey explores the intersection of narrative theory and large language models (LLMs) for automatic story generation and understandi...

[2602.15849] Preference Optimization for Review Question Generation Improves Writing Quality

This article presents IntelliReward, a novel model for generating review questions that enhances writing quality by aligning with human p...

[2602.15847] Do Personality Traits Interfere? Geometric Limitations of Steering in Large Language Models

This article explores the geometric limitations of steering personality traits in large language models (LLMs), revealing that traits are...