D4C: Improving Negative Example Quality to Enhance Machine Abstract Reasoning Ability

TL;DR

The paper tackles the challenge of abstract reasoning in AI by developing Lico-Net, a probabilistic-concept discriminator for RPM tasks, and Lico-Net-Bongard for Bongard-Logo problems. It identifies negative sample quality as a core driver of concept learning mismatches between deep models and human reasoning, and introduces two strategies, D3C and D4C, to generate higher quality negative examples. Empirical results across RPM, Bongard-Logo, and NICO demonstrate that improved negative sampling and probabilistic concept representations yield meaningful gains, with D4C often providing the strongest improvements. The work highlights the importance of negative-sample design and probabilistic reasoning as a promising direction for advancing AI abstract reasoning in vision tasks and beyond.

Abstract

This paper is dedicated to addressing the challenge of enhancing the abstract reasoning capabilities of AI, particularly for tasks involving complex human concepts. We introduce Lico-Net, a novel reasoning engine grounded in deep learning theory, which encodes the logical structure of Raven's Progressive Matrices (RPM) problems into probabilistic representations. Lico-Net excels in solving RPM tasks. Furthermore, we propose Lico-Net-Bongard, a tailored version of Lico-Net for the Bongard-Logo problem, which also achieves high reasoning accuracy through probabilistic representations. However, we observe a mismatch between the way deep learning algorithms and humans induce reasoning concepts, primarily attributed to the inadequate quality of negative samples. Improper configuration of negative samples can convey erroneous conceptual information to deep learning algorithms, thereby distorting their learning objectives. To address this issue, we propose two novel approaches: first, treating different sample points within reasoning problems as mutual negative samples to alter the existing negative sample structure in the data; second, designing a negative sample generator based on a step-wise linear attention mechanism to produce high-quality negative samples. Experimental results demonstrate that these methods significantly improve the performance of Lico-Net (-Bongard) and other baseline models on the RPM and Bongard-Logo datasets, as well as in the domain of foundational vision model processing, particularly when addressing the NICO dataset's distribution shift problem. Our findings emphasize the importance of improving negative sample quality for enhancing the abstract reasoning capabilities of deep learning algorithms and suggest that systems represent a promising direction for future research in this field.

Figures (15)

• Figure 1: RAVEN and PGM case
• Figure 2: Bongard-Logo case
• Figure 3: The figure shows the example of NICO-animal and NICO-vehicle.
• Figure 4: The figure shows the framework of Lico-Net.
• Figure 5: Perceptron of Lico-Net