Multi-modal Large Language Models (MLLMs) have exhibited impressive capability. However, recently many deficiencies of MLLMs have been found compared to human intelligence, e.g., hallucination. To drive the MLLMs study, the community dedicated efforts to building larger benchmarks with complex tasks. In this paper, we propose benchmarking an essential but usually overlooked intelligence: association, a human's basic capability to link observation and prior practice memory. To comprehensively investigate MLLM's performance on the association, we formulate the association task and devise a standard benchmark based on adjective and verb semantic concepts. Instead of costly data annotation and curation, we propose a convenient annotation-free construction method transforming the general dataset for our association tasks. Simultaneously, we devise a rigorous data refinement process to eliminate confusion in the raw dataset. Building on this database, we establish three levels of association tasks: singlestep, synchronous, and asynchronous associations. Moreover, we conduct a comprehensive investigation into the MLLMs' zero-shot association capabilities, addressing multiple dimensions, including three distinct memory strategies, both open-source and closed-source MLLMs, cutting-edge Mixture-of-Experts (MoE) models, and the involvement of human experts. Our systematic investigation shows that current open-source MLLMs consistently exhibit poor capability in our association tasks, even the currently state-of-the-art GPT-4V(vision) also has a significant gap compared to humans. We believe our benchmark would pave the way for future MLLM studies. Multi-modal Large Language Models (MLLMs) have recently made significant breakthroughs in perceiving diverse modality input and solving a broad range of tasks Zhang et al. (2024a); Carolan et al. (2024). As GPT-4V(ision) Achiam et al. (2023) and Gemini Team et al. (2023); Reid et al. (2024) address challenges that researchers have been exploring for a considerable period. Subsequently, numerous researchers have developed diverse open-source MLLMs AI et al. (2024); Bai et al. (2023b); Wang et al. (2024b); Dong et al. (2024); Liu et al. (2023a); Li et al. (2024a); Ye et al. (2023; 2024). These models usually use the Large Language Model (LLM) as the core component and expand to multi-modal with a specific module Yin et al. (2023) that transfers multi-modal tokens into language tokens, achieving alignment between different modality encoders. MLLMs demonstrated ability in visual reasoning, which requires understanding the input query and then making judgments based on the visual content. Much prior work has been dedicated to evaluating the levels of their visual reasoning capabilities. However, to the best of our knowledge, how to evaluate the association ability of MLLMs is overlooked.

Large Language Models (LLMs) have the capacity to store and recall facts. Through experimentation with open-source models, we observe that this ability to retrieve facts can be easily manipulated by changing contexts, even without altering their factual meanings. These findings highlight that LLMs might behave like an associative memory model where certain tokens in the contexts serve as clues to retrieving facts. We mathematically explore this property by studying how transformers, the building blocks of LLMs, can complete such memory tasks. We study a simple latent concept association problem with a one-layer transformer and we show theoretically and empirically that the transformer gathers information using self-attention and uses the value matrix for associative memory.

Classification models learn to generalize the associations between data samples and their target classes. However, researchers have increasingly observed that machine learning practice easily leads to systematic errors in AI applications, a phenomenon referred to as AI blindspots. Such blindspots arise when a model is trained with training samples (e.g., cat/dog classification) where important patterns (e.g., black cats) are missing or periphery/undesirable patterns (e.g., dogs with grass background) are misleading towards a certain class. Even more sophisticated techniques cannot guarantee to capture, reason about, and prevent the spurious associations. In this work, we propose ESCAPE, a visual analytic system that promotes a human-in-the-loop workflow for countering systematic errors. By allowing human users to easily inspect spurious associations, the system facilitates users to spontaneously recognize concepts associated misclassifications and evaluate mitigation strategies that can reduce biased associations. We also propose two statistical approaches, relative concept association to better quantify the associations between a concept and instances, and debias method to mitigate spurious associations. We demonstrate the utility of our proposed ESCAPE system and statistical measures through extensive evaluation including quantitative experiments, usage scenarios, expert interviews, and controlled user experiments.

Traditional cross-modal retrieval assumes explicit association of concepts across modalities, where there is no ambiguity in how the concepts are linked to each other, e.g., when we do the image search with a query "dogs", we expect to see dog images. In this paper, we consider a different setting for cross-modal retrieval where data from different modalities are implicitly linked via concepts that must be inferred by high-level reasoning; we call this setting implicit concept association. To foster future research in this setting, we present a new dataset containing 47K pairs of animated GIFs and sentences crawled from the web, in which the GIFs depict physical or emotional reactions to the scenarios described in the text (called "reaction GIFs"). We report on a user study showing that, despite the presence of implicit concept association, humans are able to identify video-sentence pairs with matching concepts, suggesting the feasibility of our task. Furthermore, we propose a novel visual-semantic embedding network based on multiple instance learning. Unlike traditional approaches, we compute multiple embeddings from each modality, each representing different concepts, and measure their similarity by considering all possible combinations of visual-semantic embeddings in the framework of multiple instance learning. We evaluate our approach on two video-sentence datasets with explicit and implicit concept association and report competitive results compared to existing approaches on cross-modal retrieval.