We generated model metamers for natural stimuli by performing gradient descent on a noise signal, matching the responses of individual layers of image and audio networks to a natural image or speech signal. The resulting signals reflect the invariances instantiated in the network up to the matched layer. We then measured whether model metamers were recognizable to human observers - a necessary condition for the model representations to replicate those of humans.

We will include another sentence in the discussion explicitly addressing the relationship to [41].

In recent years, deep neural networks have been extensively employed in perceptual systems to learn representations endowed with invariances, aiming to emulate the invariance mechanisms observed in the human brain. However, studies in the visual and auditory domains have confirmed that significant gaps remain between the invariance properties of artificial neural networks and those of humans. To investigate the invariance behavior within graph neural networks (GNNs), we introduce a model ``metamers'' generation technique. By optimizing input graphs such that their internal node activations match those of a reference graph, we obtain graphs that are equivalent in the model's representation space, yet differ significantly in both structure and node features. Our theoretical analysis focuses on two aspects: the local metamer dimension for a single node and the activation-induced volume change of the metamer manifold. Utilizing this approach, we uncover extreme levels of representational invariance across several classic GNN architectures. Although targeted modifications to model architecture and training strategies can partially mitigate this excessive invariance, they fail to fundamentally bridge the gap to human-like invariance. Finally, we quantify the deviation between metamer graphs and their original counterparts, revealing unique failure modes of current GNNs and providing a complementary benchmark for model evaluation.

Neural Information Processing Systems http://nips.cc/

We will include another sentence in the discussion explicitly addressing the relationship to [41].

Alignment between human brain networks and artificial models is actively studied in machine learning and neuroscience. A widely adopted approach to explore their functional alignment is to identify metamers for both humans and models. Metamers refer to input stimuli that are physically different but equivalent within a given system. If a model's metameric space completely matched the human metameric space, the model would achieve functional alignment with humans. However, conventional methods lack direct ways to search for human metamers. Instead, researchers first develop biologically inspired models and then infer about human metamers indirectly by testing whether model metamers also appear as metamers to humans. Here, we propose the Multidimensional Adaptive Metamer Exploration (MAME) framework, enabling direct high-dimensional exploration of human metameric space. MAME leverages online image generation guided by human perceptual feedback. Specifically, it modulates reference images across multiple dimensions by leveraging hierarchical responses from convolutional neural networks (CNNs). Generated images are presented to participants whose perceptual discriminability is assessed in a behavioral task. Based on participants' responses, subsequent image generation parameters are adaptively updated online. Using our MAME framework, we successfully measured a human metameric space of over fifty dimensions within a single experiment. Experimental results showed that human discrimination sensitivity was lower for metameric images based on low-level features compared to high-level features, which image contrast metrics could not explain. The finding suggests that the model computes low-level information not essential for human perception. Our framework has the potential to contribute to developing interpretable AI and understanding of brain function in neuroscience.

My review for this work remains the same following author response because I believe that the authors have demonstrated this work to be of high quality and relevance to the NeurIPS community. Originality: Although the algorithms used to synthesize the metamers themselves are nothing new, the work is a novel combination of previous approaches and techniques, and the analysis approach gives these methods a fresh perspective that leads to good insights. Quality: The work is of high quality and is a complete piece of work that will advance our understanding of the relationships between architecture, task and training in determining representational similarity between networks and humans (as well as between networks). Clarity: The paper is overall clear, though some details could use a bit of clarifying (what was the threshold for satisfactory termination of synthesis? Significance: This work builds on theoretical and experimental work from neuroscience used to analyze how well models of perceptual systems capture the representation within the human brain by synthesizing stimuli that match the responses of some part of the model completely and using human subjects to validate that the original and matched stimulus are in fact the same.

Dear authors, congrats on the acceptance-- this paper was discussed extensively, the the reviewers provided multiple comments and feedback-- please do take the feedback in the reviews into account when preparing your final manuscript.

In the winter of 2011, Daniel Yamins, a postdoctoral researcher in computational neuroscience at the Massachusetts Institute of Technology, would at times toil past midnight on his machine vision project. He was painstakingly designing a system that could recognize objects in pictures, regardless of variations in size, position and other properties -- something that humans do with ease. The system was a deep neural network, a type of computational device inspired by the neurological wiring of living brains. "I remember very distinctly the time when we found a neural network that actually solved the task," he said. It was 2 a.m., a tad too early to wake up his adviser, James DiCarlo, or other colleagues, so an excited Yamins took a walk in the cold Cambridge air. "I was really pumped," he said. It would have counted as a noteworthy accomplishment in artificial intelligence alone, one of many that would make neural networks the darlings of AI technology over the next few years.

In the winter of 2011, Daniel Yamins, a postdoctoral researcher in computational neuroscience at the Massachusetts Institute of Technology, would at times toil past midnight on his machine vision project. He was painstakingly designing a system that could recognize objects in pictures, regardless of variations in size, position and other properties -- something that humans do with ease. The system was a deep neural network, a type of computational device inspired by the neurological wiring of living brains. "I remember very distinctly the time when we found a neural network that actually solved the task," he said. It was 2 a.m., a tad too early to wake up his adviser, James DiCarlo, or other colleagues, so an excited Yamins took a walk in the cold Cambridge air. "I was really pumped," he said. It would have counted as a noteworthy accomplishment in artificial intelligence alone, one of many that would make neural networks the darlings of AI technology over the next few years.