6 papers
Sequential Enumeration in Large Language Models
Kuinan Hou, Marco Zorzi, Alberto Testolin
Reliably counting and generating sequences of items remain a significant challenge for neural networks, including Large Language Models (LLMs). Indeed, although this capability is…
A Rate-Distortion Perspective on the Emergence of Number Sense in Unsupervised Generative Models
Leo D'Amato, Davide Nuzzi, Alberto Testolin +3
Number sense is a core cognitive ability supporting various adaptive behaviors and is foundational for mathematical learning. Here, we study its emergence in unsupervised generativ…
Assessing the Visual Enumeration Abilities of Specialized Counting Architectures and Vision-Language Models
Kuinan Hou, Jing Mi, Marco Zorzi +2
Counting the number of items in a visual scene remains a fundamental yet challenging task in computer vision. Traditional approaches to solving this problem rely on domain-specific…
Visual Enumeration Remains Challenging for Multimodal Generative AI
Alberto Testolin, Kuinan Hou, Marco Zorzi
Many animal species can approximately judge the number of objects in a visual scene at a single glance, and humans can further determine the exact cardinality of a set by deploying…
Learning neuro-symbolic convergent term rewriting systems
Flavio Petruzzellis, Alberto Testolin, Alessandro Sperduti
Building neural systems that can learn to execute symbolic algorithms is a challenging open problem in artificial intelligence, especially when aiming for strong generalization and…
Artificial Intelligence Can Emulate Human Normative Judgments on Emotional Visual Scenes
Zaira Romeo, Alberto Testolin
Affective reactions have deep biological foundations, however in humans the development of emotion concepts is also shaped by language and higher-order cognition. A recent breakthr…