OFFSIDE: Benchmarking Unlearning Misinformation in Multimodal Large Language Models

Recent advancements in out-of-context (OOC) misinformation detection have highlighted the need for improved consistency checks between image-text pairs and external evidence. The proposed HiEAG framework aims to enhance this process by utilizing multimodal large language models (MLLMs) to refine external consistency checking. This approach includes a comprehensive pipeline that integrates evidence reranking and rewriting, addressing the limitations of current methods that focus primarily on internal consistency.

Vision-Language Models (VLMs) are emerging models that integrate visual content with natural language. Current methods typically fuse data either early in the encoding process or late through pooled embeddings. This paper introduces a lightweight fusion module utilizing cross-only, bidirectional attention layers to align hidden states from both modalities, enhancing understanding while keeping encoders non-causal. The proposed method aims to improve the performance of VLMs by leveraging the inherent structure of visual and textual data.

The paper titled 'Unifying Segment Anything in Microscopy with Vision-Language Knowledge' discusses the importance of accurate segmentation in biomedical images. It highlights the limitations of existing models in handling unseen domain data due to a lack of vision-language knowledge. The authors propose a new framework, uLLSAM, which utilizes Multimodal Large Language Models (MLLMs) to enhance segmentation performance. This approach aims to improve generalization capabilities across cross-domain datasets, achieving notable performance improvements.

The paper titled 'Bias-Restrained Prefix Representation Finetuning for Mathematical Reasoning' introduces a new method called Bias-REstrained Prefix Representation FineTuning (BREP ReFT). This approach aims to enhance the mathematical reasoning capabilities of models by addressing the limitations of existing Representation finetuning (ReFT) methods, which struggle with mathematical tasks. The study demonstrates that BREP ReFT outperforms both standard ReFT and weight-based Parameter-Efficient finetuning (PEFT) methods through extensive experiments.

The study investigates the ability of transformer models to predict long steps in the Collatz sequence, a complex arithmetic function that maps odd integers to their successors. The accuracy of the models varies significantly depending on the base used for encoding, achieving up to 99.7% accuracy for bases 24 and 32, while dropping to 37% and 25% for bases 11 and 3. Despite these variations, all models exhibit a common learning pattern, accurately predicting inputs with similar residuals modulo 2^p.

Higher-order Neural Additive Models (HONAMs) have been introduced as an advancement over Neural Additive Models (NAMs), which are known for their predictive performance and interpretability. HONAMs address the limitation of NAMs by effectively capturing feature interactions of arbitrary orders, enhancing predictive accuracy while maintaining interpretability, crucial for high-stakes applications. The source code for HONAM is publicly available on GitHub.