83 articles tagged with #dataset. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Researchers introduced D-Negation, a new dataset and learning framework that improves vision-language AI models' ability to understand negative semantics and complex expressions. The approach achieved up to 5.7 mAP improvement on negative semantic evaluations while fine-tuning less than 10% of model parameters.
Researchers introduce a new framework showing that emotional tone in text systematically affects how large language models process and reason over information. They developed AURA-QA, an emotionally balanced dataset, and proposed emotional regularization techniques that improve reading comprehension performance across multiple benchmarks.
Researchers introduce ARAS400k, a large-scale remote sensing dataset containing 400k images (100k real, 300k synthetic) with segmentation maps and descriptions. The study demonstrates that combining real and synthetic data consistently outperforms training on real data alone for semantic segmentation and image captioning tasks.
Researchers introduce Answer-Then-Check, a novel safety alignment approach for large language models that enables them to evaluate response safety before outputting to users. The method uses a new 80K-sample dataset called Reasoned Safety Alignment (ReSA) and demonstrates improved jailbreak defense while maintaining general reasoning capabilities.
Researchers have released Tucano 2, an open-source suite of Portuguese language models ranging from 0.5-3.7 billion parameters, featuring enhanced datasets and training recipes. The models achieve state-of-the-art performance on Portuguese benchmarks and include capabilities for coding, tool use, and chain-of-thought reasoning.
Researchers introduce CHIMERA, a compact 9K-sample synthetic dataset that enables smaller AI models to achieve reasoning performance comparable to much larger models. The dataset addresses key challenges in training reasoning-capable LLMs through automated generation and cross-validation across 8 scientific disciplines.
Researchers introduced InterSyn, a 1.8M sample dataset designed to improve Large Multimodal Models' ability to generate interleaved image-text content. The dataset includes a new evaluation framework called SynJudge that measures four key performance metrics, with experiments showing significant improvements even with smaller 25K-50K sample subsets.
Researchers introduce EgoNight, the first comprehensive benchmark for nighttime egocentric vision understanding, featuring day-night aligned videos and visual question answering tasks. The benchmark reveals significant performance drops in state-of-the-art multimodal large language models when operating under low-light conditions.
Researchers conducted the first comprehensive analysis of open-source direct preference optimization (DPO) datasets used to align large language models, revealing significant quality variations. They created UltraMix, a curated dataset that's 30% smaller than existing options while delivering superior performance across benchmarks.
Researchers have released HumanMCP, the first large-scale dataset designed to evaluate tool retrieval performance in Model Context Protocol (MCP) servers. The dataset addresses a critical gap by providing realistic, human-like queries paired with 2,800 tools across 308 MCP servers, improving upon existing benchmarks that lack authentic user interaction patterns.
Researchers introduce DesignSense-10k, a dataset of 10,235 human-annotated preference pairs for evaluating graphic layout generation, along with DesignSense, a specialized AI model that outperforms existing models by 54.6% in layout quality assessment. The framework addresses the gap between AI-generated layouts and human aesthetic preferences, showing practical improvements in layout generation through reinforcement learning.
Researchers released LFQA-HP-1M, a dataset with 1.3 million human preference annotations for evaluating long-form question answering systems. The study introduces nine quality rubrics and shows that simple linear models can match advanced LLM evaluators while exposing vulnerabilities in current evaluation methods.
Researchers developed Lipi-Ghor-882, an 882-hour Bengali speech dataset, and demonstrated that targeted fine-tuning with synthetic acoustic degradation significantly improves automatic speech recognition for long-form Bengali audio. Their dual pipeline achieved a 0.019 Real-Time Factor, establishing new benchmarks for low-resource speech processing.
Researchers released the Asta Interaction Dataset containing over 200,000 user queries from AI-powered scientific research tools, revealing how scientists interact with LLM-based research assistants. The study shows users treat these systems as collaborative research partners, submitting longer queries and using outputs as persistent artifacts for non-linear exploration.
The article title suggests LeRobot has released the world's largest open-source self-driving dataset, representing a significant contribution to autonomous vehicle development. However, no article body was provided for detailed analysis.
Researchers introduced the AgrI Challenge, a data-centric AI competition focused on agricultural vision that revealed significant generalization gaps in machine learning models when deployed across different field conditions. The study found that models trained on single datasets showed validation-test gaps of up to 16.20%, but collaborative multi-source training reduced these gaps to under 3%.
Researchers introduced the Contextual Emotional Inference (CEI) Benchmark, a dataset of 300 human-validated scenarios designed to evaluate how well large language models understand pragmatic reasoning in complex communication. The benchmark tests LLMs' ability to interpret ambiguous utterances across five pragmatic subtypes including sarcasm, mixed signals, and passive aggression in various social contexts.
Researchers developed a methodology to fine-tune large language models (LLMs) for generating code-switched text between English and Spanish by back-translating natural code-switched sentences into monolingual English. The study found that fine-tuning significantly improves LLMs' ability to generate fluent code-switched text, and that LLM-based evaluation methods align better with human preferences than traditional metrics.
Researchers have created a new multi-task Chinese dialogue dataset that enables prediction of user satisfaction, emotion recognition, and emotional state transitions across multiple conversation turns. The dataset addresses limitations in existing Chinese resources and aims to improve understanding of how user emotions evolve during interactions to better predict satisfaction.
Researchers have released MuRAL, a new dataset containing over 21 hours of multi-resident smart home sensor data with natural language annotations for training AI models. The dataset aims to improve Large Language Models' ability to understand human activities in complex smart home environments, though current LLMs still struggle with key tasks like resident identification and activity prediction.
Researchers introduce CareMedEval, a new dataset with 534 questions based on 37 scientific articles to evaluate large language models' ability to perform critical appraisal in biomedical contexts. Testing reveals current AI models struggle with this specialized reasoning task, achieving only 0.5 exact match rates even with advanced prompting techniques.
Researchers have released the Vienna 4G/5G Drive-Test Dataset, a comprehensive open dataset of georeferenced mobile network measurements collected across Vienna, Austria. The dataset combines passive scanner observations with active handset logs and includes building/terrain models to support machine learning applications in mobile network analysis and optimization.
Researchers have created CrimeNER, a specialized dataset of over 1,500 annotated crime-related documents for training named-entity recognition AI models. The study addresses the lack of quality training data in the crime domain by developing a database from terrorist attack reports and DOJ press notes, defining 22 types of crime-related entities.
Researchers have created MAC, the first public conversion rate prediction dataset featuring labels from multiple attribution mechanisms, along with PyMAL, an open-source library for multi-attribution learning approaches. The study introduces a new method called Mixture of Asymmetric Experts (MoAE) that significantly outperforms existing state-of-the-art multi-attribution learning methods.
The article appears to be about an open preference dataset for text-to-image generation created by the Hugging Face community. However, the article body is empty, making it impossible to provide specific details about the dataset's features, applications, or significance.
