Blocks as Probes: Dissecting Categorization Ability of Large Multimodal Models

• URL: http://arxiv.org/abs/2409.01560v1
• Date: Tue, 3 Sep 2024 02:55:36 GMT
• Title: Blocks as Probes: Dissecting Categorization Ability of Large Multimodal Models
• Authors: Bin Fu, Qiyang Wan, Jialin Li, Ruiping Wang, Xilin Chen,
• Abstract summary: Recent development of Large Multimodal Models (LMMs) has demonstrated impressive results in high-level visual tasks. We propose a novel, challenging, and efficient benchmark based on composite blocks, called ComBo, which provides a disentangled evaluation framework. We find that although LMMs exhibit acceptable generalization ability in learning new categories, there are still gaps compared to humans in many ways.
• Score: 31.47100708645748
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Categorization, a core cognitive ability in humans that organizes objects based on common features, is essential to cognitive science as well as computer vision. To evaluate the categorization ability of visual AI models, various proxy tasks on recognition from datasets to open world scenarios have been proposed. Recent development of Large Multimodal Models (LMMs) has demonstrated impressive results in high-level visual tasks, such as visual question answering, video temporal reasoning, etc., utilizing the advanced architectures and large-scale multimodal instruction tuning. Previous researchers have developed holistic benchmarks to measure the high-level visual capability of LMMs, but there is still a lack of pure and in-depth quantitative evaluation of the most fundamental categorization ability. According to the research on human cognitive process, categorization can be seen as including two parts: category learning and category use. Inspired by this, we propose a novel, challenging, and efficient benchmark based on composite blocks, called ComBo, which provides a disentangled evaluation framework and covers the entire categorization process from learning to use. By analyzing the results of multiple evaluation tasks, we find that although LMMs exhibit acceptable generalization ability in learning new categories, there are still gaps compared to humans in many ways, such as fine-grained perception of spatial relationship and abstract category understanding. Through the study of categorization, we can provide inspiration for the further development of LMMs in terms of interpretability and generalization.

Related papers

• Contextuality Helps Representation Learning for Generalized Category Discovery [5.885208652383516]
  This paper introduces a novel approach to Generalized Category Discovery (GCD) by leveraging the concept of contextuality. Our model integrates two levels of contextuality: instance-level, where nearest-neighbor contexts are utilized for contrastive learning, and cluster-level, employing contrastive learning. The integration of the contextual information effectively improves the feature learning and thereby the classification accuracy of all categories.
  arXiv  Detail & Related papers  (2024-07-29T07:30:41Z)
• Coding for Intelligence from the Perspective of Category [66.14012258680992]
  Coding targets compressing and reconstructing data, and intelligence. Recent trends demonstrate the potential homogeneity of these two fields. We propose a novel problem of Coding for Intelligence from the category theory view.
  arXiv  Detail & Related papers  (2024-07-01T07:05:44Z)
• Aligning Vision Models with Human Aesthetics in Retrieval: Benchmarks and Algorithms [91.19304518033144]
  We aim to align vision models with human aesthetic standards in a retrieval system. We propose a preference-based reinforcement learning method that fine-tunes the vision models to better align the vision models with human aesthetics.
  arXiv  Detail & Related papers  (2024-06-13T17:59:20Z)
• VisualCritic: Making LMMs Perceive Visual Quality Like Humans [65.59779450136399]
  We present VisualCritic, the first LMM for broad-spectrum image subjective quality assessment. VisualCritic can be used across diverse data right out of box, without any requirements of dataset-specific adaptation operations.
  arXiv  Detail & Related papers  (2024-03-19T15:07:08Z)
• Unsupervised discovery of Interpretable Visual Concepts [0.0]
  We propose two methods to explain a model's decision, enhancing global interpretability. One method is inspired by Occlusion and Sensitivity analysis (incorporating causality) The other method uses a novel metric, called Class-aware Order Correlation (CaOC), to globally evaluate the most important image regions.
  arXiv  Detail & Related papers  (2023-08-31T07:53:02Z)
• Semantic Representation and Dependency Learning for Multi-Label Image Recognition [76.52120002993728]
  We propose a novel and effective semantic representation and dependency learning (SRDL) framework to learn category-specific semantic representation for each category. Specifically, we design a category-specific attentional regions (CAR) module to generate channel/spatial-wise attention matrices to guide model. We also design an object erasing (OE) module to implicitly learn semantic dependency among categories by erasing semantic-aware regions.
  arXiv  Detail & Related papers  (2022-04-08T00:55:15Z)
• A Minimalist Dataset for Systematic Generalization of Perception, Syntax, and Semantics [131.93113552146195]
  We present a new dataset, Handwritten arithmetic with INTegers (HINT), to examine machines' capability of learning generalizable concepts. In HINT, machines are tasked with learning how concepts are perceived from raw signals such as images. We undertake extensive experiments with various sequence-to-sequence models, including RNNs, Transformers, and GPT-3.
  arXiv  Detail & Related papers  (2021-03-02T01:32:54Z)
• Enhancing efficiency of object recognition in different categorization levels by reinforcement learning in modular spiking neural networks [1.392250707100996]
  We propose a computational model for object recognition in different categorization levels. A spiking neural network equipped with the reinforcement learning rule is used as a module at each categorization level. According to the required information at each categorization level, the relevant band-pass filtered images are utilized.
  arXiv  Detail & Related papers  (2021-02-10T12:33:20Z)
• Concept Learners for Few-Shot Learning [76.08585517480807]
  We propose COMET, a meta-learning method that improves generalization ability by learning to learn along human-interpretable concept dimensions. We evaluate our model on few-shot tasks from diverse domains, including fine-grained image classification, document categorization and cell type annotation.
  arXiv  Detail & Related papers  (2020-07-14T22:04:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.