​[[{“value”:”

Humans excel at processing vast arrays of visual information, a skill that is crucial for achieving artificial general intelligence (AGI). Over the decades, AI researchers have developed Visual Question Answering (VQA) systems to interpret scenes within single images and answer related questions. While recent advancements in foundation models have significantly closed the gap between human and machine visual processing, conventional VQA has been restricted to reason about only single images at a time rather than whole collections of visual data.

This limitation poses challenges in more complex scenarios. Take, for example, the challenges of discerning patterns in collections of medical images, monitoring deforestation through satellite imagery, mapping urban changes using autonomous navigation data, analyzing thematic elements across large art collections, or understanding consumer behavior from retail surveillance footage. Each of these scenarios entails not only visual processing across hundreds or thousands of images but also necessitates cross-image processing of these findings. To address this gap, this project focuses on the “Multi-Image Question Answering” (MIQA) task, which exceeds the reach of traditional VQA systems.

Visual Haystacks: the first “visual-centric” Needle-In-A-Haystack (NIAH) benchmark designed to rigorously evaluate Large Multimodal Models (LMMs) in processing long-context visual information.

Read More »Are We Ready for Multi-Image Reasoning? Launching VHs: The Visual Haystacks Benchmark! The Berkeley Artificial Intelligence Research Blog

​[[{“value”:” Large language models (LLMs) have showcased remarkable capabilities in generating content and solving complex problems across various domains. However, a notable challenge persists in their ability to perform multi-step deductive reasoning. This type of reasoning requires a coherent and logical thought process over extended… Read More »Researchers from the University of Auckland Introduced ChatLogic: Enhancing Multi-Step Reasoning in Large Language Models with Over 50% Accuracy Improvement in Complex Tasks Sana Hassan Artificial Intelligence Category – MarkTechPost

​[[{“value”:” Using offline web apps and AI apps often comes with challenges. Users typically need to navigate multiple steps to get an app running. These steps can be confusing and time-consuming, especially for those who are not tech-savvy. Additionally, managing and customizing these apps often… Read More »Pinokio 2.0: A New Pinokio Browser Version that Lets You Locally Install, Run, and Automate Any AI on Your Computer Niharika Singh Artificial Intelligence Category – MarkTechPost

​[[{“value”:” Evaluating the retrieval and reasoning capabilities of large language models (LLMs) in extremely long contexts, extending up to 1 million tokens, is a significant challenge. Efficiently processing long texts is crucial for extracting relevant information and making accurate decisions based on extensive data. This… Read More »NeedleBench: A Customizable Dataset Framework that Includes Tasks for Evaluating the Bilingual Long-Context Capabilities of LLMs Across Multiple Length Intervals Aswin Ak Artificial Intelligence Category – MarkTechPost

​[[{“value”:” Despite their expanding capabilities, large language models (LLMs) need help with processing extensive contexts. These limitations stem from Transformer-based architectures struggling to extrapolate beyond their training window size. Processing long token sequences requires substantial computational resources and risks producing noisy attention embeddings. These constraints… Read More »EM-LLM: A Novel and Flexible Architecture that Integrates Key Aspects of Human Episodic Memory and Event Cognition into Transformer-based Language Models Mohammad Asjad Artificial Intelligence Category – MarkTechPost

​[[{“value”:” Innovation and the artistic, musical, and literary expression of human experiences and emotions depend on creativity. However, the idea that material created by humans is inherently better is coming under pressure from the emergence of generative artificial intelligence (AI) technologies, such as Large Language… Read More »Is Generative AI Boosting Individual Creativity but  Reducing Collective Novelty? Tanya Malhotra Artificial Intelligence Category – MarkTechPost

​[[{“value”:” LLMs excel in natural language processing tasks but face deployment challenges due to high computational and memory demands during inference. Recent research [MWM+24, WMD+23, SXZ+24, XGZC23, LKM23] aims to enhance LLM efficiency through quantization, pruning, distillation, and improved decoding. Sparsity, a key approach, reduces… Read More »Q-Sparse: A New Artificial Intelligence AI Approach to Enable Full Sparsity of Activations in LLMs Sana Hassan Artificial Intelligence Category – MarkTechPost

​[[{“value”:” Snowflake recently announced the release of its updated text embedding model, snowflake-arctic-embed-m-v1.5. This model generates highly compressible embedding vectors while maintaining high performance. The model’s most noteworthy feature is its ability to produce embedding vectors compressed to as small as 128 bytes per vector… Read More »Snowflake-Arctic-Embed-m-v1.5 Released: A 109M Parameters Groundbreaking Text Embedding Model with Enhanced Compression and Performance Capabilities Asif Razzaq Artificial Intelligence Category – MarkTechPost

​[[{“value”:” Large Language Models (LLMs) and their multi-modal counterparts (MLLMs) have made significant strides in advancing artificial general intelligence (AGI) across various domains. However, these models face a significant challenge in the realm of visual mathematical problem-solving. While MLLMs have demonstrated impressive capabilities in diverse… Read More »From Diagrams to Solutions: MAVIS’s Three-Stage Framework for Mathematical AI Mohammad Asjad Artificial Intelligence Category – MarkTechPost