Learning Multimodal Embeddings for Traffic Accident Prediction and Causal Estimation

A new paper on arXiv presents datasets aimed at improving the understanding of idiomatic and figurative language in Natural Language Processing (NLP). These datasets are designed to assist large language models (LLMs) in better interpreting informal language, which has become increasingly prevalent in social media and everyday communication.

A novel self-supervised symbolic auto-encoder has been introduced, enabling symbolic computer vision to interpret diagrams through structured representations and logical rules. This approach contrasts with traditional pixel-based visual models by parsing diagrams into geometric primitives, enhancing machine vision's interpretability.

FloodDiffusion has been introduced as a novel framework for text-driven, streaming human motion generation, capable of producing seamless motion sequences in real-time based on time-varying text prompts. This approach improves upon existing methods by employing a tailored diffusion forcing framework that addresses the limitations of traditional models, ensuring better alignment with real motion distributions.

A new method for robust multimodal sentiment analysis of image-text pairs has been proposed, addressing challenges related to low-quality and missing modalities. The Distribution-based feature Recovery and Fusion (DRF) technique utilizes a feature queue for each modality to approximate feature distributions, enhancing sentiment prediction accuracy in real-world applications.

A new method for low-light image denoising has been proposed, which requires minimal data acquisition by synthesizing noise from a single noisy image and a dark frame per ISO setting. This approach utilizes a Poisson distribution to model signal-dependent noise and a Fourier-domain spectral sampling algorithm for signal-independent noise, aiming to improve image quality in challenging lighting conditions.

FireSentry has been introduced as a multi-modal dataset designed for fine-grained wildfire spread forecasting, utilizing sub-meter spatial and sub-second temporal resolution data collected via UAV platforms. This dataset includes visible and infrared video streams, environmental measurements, and validated fire masks, addressing the limitations of existing coarse-scale models that rely on low-resolution satellite data.

A new framework has been introduced for identifying attributions of causality in political text, utilizing a lightweight causal language model to generate structured data sets of causal claims. This approach aims to enhance the systematic analysis of explanations in political science, an area that has been historically fragmented and underdeveloped.

A recent study introduces a group fairness lens for evaluating large language models (LLMs), proposing a novel hierarchical schema to assess bias and fairness. The research presents the GFAIR dataset and introduces GF-THINK, a method aimed at mitigating biases in LLMs, highlighting the critical need for broader evaluations of these models beyond traditional metrics.