All notes, newest first
2,343 notes. Notes 1,051 to 1,100, page 22 of 47.
Disentangling Curriculum Learning in NLP: Towards a Unifying Taxonomy
The paper proposes a taxonomy to better understand curriculum learning (CL) in Natural Language Processing (NLP). The taxonomy aims to provide a systematic analysis of CL strategies by separating difficulty evaluation…
Find Before You Fine-Tune: A Diagnostic Study of Small LLMs for Cybersecurity QA
The paper presents a diagnostic framework called FiT (Find before Fine-Tune) to evaluate the suitability of small language models for cybersecurity question-answering tasks. The study finds that fine-tuning these…
H$^2$SD: Hybrid Hindsight Self-Distillation
The paper introduces H$^2$SD, a framework for reinforcement learning with verifiable rewards (RLVR) that addresses the limitations of previous methods. H$^2$SD uses a hybrid approach to self-distillation, where it…
Is EEG-to-Text Feasible in Real-World Scenarios? An In-Depth Analysis Using a Neuropsychology-Inspired Benchmark
Researchers have created a new benchmark called Corpus OF Eeg-To-Text (COFETT) to evaluate the feasibility of translating brain signals into text using Electroencephalography (EEG) in real-world scenarios. The study…
Measuring AI innovation with trademark data
This research note discusses the potential of using trademark data to measure AI innovation. The authors argue that AI trademarks can provide a complementary perspective to AI patents by capturing different aspects of…
Measuring Reward-Seeking via Contrastive Belief Updates
Researchers have developed a method to measure 'reward-seeking' in language models trained with reinforcement learning. This behavior occurs when the model optimizes for the grader's judgment instead of the intended…
MedDDC-Eval: Diagnosis-Decoupled Evaluation of Multi-Turn Medical Consultation Agents
Researchers have developed a new testbed called MedDDC-Eval for evaluating multi-turn medical consultation agents. This testbed separates the quality of the elicited history from the terminal diagnosis generation…
PLAID-PRF: Pseudo-Relevance Feedback with Centroid-like Tokens in PLAID
Researchers propose a method called PLAID-PRF to improve the effectiveness of multi-vector dense retrieval models like ColBERT. This method uses centroid-based quantization in PLAID, a model that reduces index size and…
RAGAL: A Frugal, Fully Local Retrieval-Augmented Assistant for Technical Support at a Government Agency
The paper introduces RAGAL, a locally-hosted AI assistant developed for a Romanian agency's technical support team. The system was designed to function within strict data locality constraints, with no external API…
Rationale-Guided Knowledge Distillation for Cross-Lingual Stance Detection
The authors propose a method called rationale-guided knowledge distillation for cross-lingual stance detection. This technique aims to identify attitudes in texts towards a target, and helps when there is a lack of…
Selection Shapes the Boundary: A Preregistered Replication of Monotonicity and Label Agreement in Unselected NLI Populations
The study aimed to replicate an earlier finding about lower label agreement in natural language inference (NLI) when using non-upward monotonicity operators, specifically in the ChaosNLI dataset. The current research…
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Researchers introduce a new method called Staleness-Adaptive Trust Region (SAT) for asynchronous reinforcement learning. SAT aims to address the issue of 'staleness' in asynchronous learning, which occurs due to policy…
Transcription Policy as a Latent Variable: Activating Controllable Verbatim ASR with Word-Level Timing
The authors propose a method to improve the accuracy of automatic speech recognition (ASR) models by controlling the activation of 'verbatim' versus 'intended' transcription styles. They argue that current models…
AdaHome: An Adaptive Smart Home Assistant using Local Small Language Models
A paper titled 'AdaHome: An Adaptive Smart Home Assistant using Local Small Language Models' presents a new smart home assistant system designed to work with locally deployed small language models in…
Adaptive Adversaries: A Multi-Turn, Multi-LLM Benchmark for LLM Agent Security
The paper presents a new benchmark for testing the security of large language model (LLM) agents called 'Adaptive Adversaries'. The benchmark involves 21 scenarios where an adaptive attacker can observe prior defender…
Adaptive Adversaries: A Multi-Turn, Multi-LLM Benchmark for LLM Agent Security
The paper presents a new benchmark for testing the security of LLM (Language Model) agents against adaptive adversaries in multi-turn interactions. The benchmark, called Adaptive Adversaries, evaluates how defenders…
Automated Discovery Has No Universally Superior Harness
A study investigates the effectiveness of different automated discovery systems (OpenEvolve and TTT-Discover) in various machine learning models and problems. The research finds that no single system consistently…
Automated Discovery Has No Universally Superior Harness
A study evaluates various discovery harnesses, such as OpenEvolve and TTT-Discover, used in automated discovery systems. The research finds that no single harness performs consistently well across different models and…
Autoresearch with Coding Agents: Generalizers and Metric-Maximizers on Quran Recitation Data
Researchers tested two coding agents, Claude Code and OpenAI Codex, on a real-world task of identifying Quranic verses in noisy speech transcripts. Both agents developed similar algorithms initially but diverged later…
Autoresearch with Coding Agents: Generalizers and Metric-Maximizers on Quran Recitation Data
Two coding agents were tested on a production task involving Quranic verse identification and transcription splitting. Both agents independently developed similar algorithms, but one (OpenAI Codex) performed poorly due…
Can We Break LLMs Out of Self-Loops? Fine-Grained Reasoning Control with Activation Steering
Researchers propose a new method called SOPHIA for controlling the reasoning process of large language models, which are currently uncontrollable despite their extended use. The proposed method involves treating each…
Chemical filters for ultra-high-throughput materials screening and generation
Researchers introduce a chemical validity operator to improve the reliability of AI-generated materials designs by filtering out implausible compositions that violate established chemical principles. The operator…
Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints
In the study titled 'Do Language Models Dream of Binding Molecules? Benchmarking LLMs under Spatial Constraints', the authors examine the ability of large language models (LLMs) to navigate complex 3D spatial…
Enhancing Rubric-based RL via Self-Distillation
The paper proposes a new method called Criterion-Distilled Policy Optimization (CriPO) to address two limitations in Rubric-based Reinforcement Learning (RL). The first limitation, Limited Exploration, occurs when…
Exploratory and Assimilating Reflection: Reflective Recall Cycle for Long-term Memory
The paper proposes a framework called Exploratory-Assimilating Reflection (EAR) for autonomous agents with limited context window and external memory. EAR aims to improve retrieval performance by combining two…
GigaPath-Flash and GigaTIME-Flash: Efficient Pathology Foundation Models for Whole-Slide and Tumor Microenvironment Analysis
Researchers have developed two new foundation models called GigaPath-Flash and GigaTIME-Flash for analyzing whole-slide pathology and predicting the tumor immune microenvironment from routine H&E images. These models…
GigaPath-Flash and GigaTIME-Flash: Efficient Pathology Foundation Models for Whole-Slide and Tumor Microenvironment Analysis
Researchers have developed two efficient deep learning models named GigaPath-Flash and GigaTIME-Flash for whole-slide pathology analysis and spatial proteomics prediction. These models are based on a compact ViT-S tile…
Harness Engineering for LLM-Driven GPU Kernel Generation
The paper discusses a system for optimizing GPU kernels using large language models (LLMs). The system separates an evaluation harness from a profile-backed optimization controller to enforce compilation, correctness…
Harness Engineering for LLM-Driven GPU Kernel Generation
This research presents a system that uses large language models (LLMs) to optimize GPU kernels. The system separates an evaluation harness from an optimization controller, with human-authored skills and code agents…
HAS: Highlight-guided Attention Steering for Multimodal LLM Video Summarization
In this research paper, the authors propose a method called Highlight-guided Attention Steering (HAS) for improving video summarization using Multimodal Large Language Models (M-LLMs). HAS aims to enhance the…
HAS: Highlight-guided Attention Steering for Multimodal LLM Video Summarization
The paper proposes a new method called Highlight-guided Attention Steering (HAS) for video summarization using Multimodal Large Language Models (M-LLMs). HAS aims to enhance the understanding of videos by focusing…
How Does Alignment Tuning Shape Representations of Sycophancy and Related Cue-Induced Biases in LLMs?
The study investigates the susceptibility of large language models (LLMs) to biases induced by subtle changes in input prompts. The research reveals that these biases are primarily shaped during alignment tuning rather…
How Does Alignment Tuning Shape Representations of Sycophancy and Related Cue-Induced Biases in LLMs?
The research investigates the susceptibility of large language models (LLMs) to biases induced by subtle changes in input prompts. The study finds that these biases are largely due to alignment tuning rather than…
Human Grounded Evaluation of Large Language Models for Optical Network Automation
The paper introduces a new evaluation pipeline named HuGLEN for comparing large language models (LLMs) used in network automation. The pipeline uses an LLM to evaluate other LLMs and rates them based on a quality…
I wanted it to feel more personal: Customization of social AI as AI individualism in practice
This research study explores why and how users customize social AI like ChatGPT, Grok, and Character.ai. The researchers found seven motivations for customization: enhanced pragmatic support, emotional support or…
Judge-dependent safety gains and model-specific helpfulness costs of evidence-sufficiency prompting in clinical LLMs
A study investigates the impact of a structured evidence-sufficiency prompt on reducing overconfident answers from four language models in a clinical context. The study finds a significant reduction in unsafe…
Judge-dependent safety gains and model-specific helpfulness costs of evidence-sufficiency prompting in clinical LLMs
A study investigates the impact of a structured evidence-sufficiency prompt on reducing unsafe overconfident answers from four different language models in a medical context. The study finds that the prompt…
LLMs and Agentic AI Systems for Smart Grids: A Tutorial on Architectures and Applications
This paper discusses how large language models (LLMs) and agentic AI systems are being used in smart grids for tasks like forecasting, optimization, and control. The authors present a design principle called the…
LLMs and Agentic AI Systems for Smart Grids: A Tutorial on Architectures and Applications
The paper discusses how large language models (LLMs) and agentic AI systems are being used in smart grids for tasks such as forecasting, optimization, and control. The authors present a design principle that ensures…
LLMs and Agentic AI Systems for Smart Grids: A Tutorial on Architectures and Applications
The paper discusses the use of Large Language Models (LLMs) and Agentic AI systems in smart grids for tasks like forecasting, optimization, and control. It presents a design principle to ensure numerical results come…
MADA-RL: Multi-Agent Debate-Aware Reinforcement Learning for Parameter-Efficient Reasoning in Compact Models
The paper introduces MADA-RL, a method for fine-tuning compact language models with limited budgets. It specializes the models into generator and critic roles and trains them using a debate-aware learning signal. The…
MADA-RL: Multi-Agent Debate-Aware Reinforcement Learning for Parameter-Efficient Reasoning in Compact Models
Researchers introduce MADA-RL, a framework that improves reasoning performance in compact models (up to 4 billion parameters) while reducing training cost. This is achieved by specializing the models into generator and…
Natural Language Access to Domain-Specific Metadata: A Reusable Framework for LLM Query Generation
Researchers have developed a system called Natural Language Knowledge Graph Query (NLKGQ) that allows researchers to ask ad-hoc questions about the contents of domain-specific archives using natural language. The…
Natural Language Access to Domain-Specific Metadata: A Reusable Framework for LLM Query Generation
Researchers have developed a system called Natural Language Knowledge Graph Query (NLKGQ) that allows them to ask ad-hoc questions about the contents of domain-specific archives using natural language, without…
O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning
Researchers have developed a new method for detecting abnormal objects and events in industrial settings using an object-centric tracking and reasoning approach. This method is designed to overcome the limitations of…
O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning
Researchers have developed a new method for Industrial Video Anomaly Detection (IVAD) that aims to identify unusual objects and events in industrial processes. This method is designed to work in complex settings where…
OntoExtend: A Framework for Requirement-driven and Scalable Ontology Extension with LLMs
The paper introduces OntoExtend, a framework that uses Large Language Models (LLMs) to extend existing ontologies based on requirements or competency questions. The framework was evaluated on two use cases and found to…
OR Else: A Differentiable Trust Region for Policy Optimization
The paper compares PPO (Proximal Policy Optimization) and GRPO (Group Regularized Policy Optimization) with two variations using a differentiable trust region for policy optimization called Output Reset (OR). The study…
PEARL: Auditable Repair for Scientific Reasoning Graph Extraction
The paper proposes a new framework called PEARL for extracting reliable scientific reasoning graphs from Large Language Models (LLMs). The LLM outputs often contain errors such as malformed syntax, incorrectly oriented…
Rethinking Heterogeneous LLM Merging: A Weighted Model Averaging Perspective
Researchers propose a simple method to merge large language models with significantly different parameters without training or alignment by using dimensional adaptation and ratio-controlled interpolation. The method…
Page 22 of 47
