FlowGenesis is a framework that lets neural networks self‑organize their architecture during training by adding or removing neurons based on a local statistic called Representation Flux. Neurons with high Flux are split to preserve function, while low‑Flux, low‑activity neurons are pruned. Experiments on image and text classification show that networks starting from different sizes converge to similar neuron counts and performance, and the method also improves continual learning by reducing parameters for new tasks without sacrificing accuracy.
By Maksim A Kazanskii
The paper argues that neural networks can exhibit modular organization even when individual neurons are not specialized. It introduces a nested hierarchy of modularity—unit-level, orthogonal, and linear—and shows that many vision, language, and RNN models lack unit-level modularity but still display clear functional modularity at the subspace level. The authors demonstrate that these subspaces can be causally validated through ablation and align with brain data, revealing distinct content/style or syntax/semantics separations.
By Shreya Saha, Benjamin Bergen, Meenakshi Khosla
arXiv:2606. 28380v1 Announce Type: cross Abstract: The intricate structures of biological neural networks largely emerge during development, guided by a comparatively compressed blueprint encoded in the genome.
By Mani Hamidi, Sina Khajehabdollahi, Charley M. Wu, Emmanouil Giannakakis
arXiv:2410. 24050v3 Announce Type: replace Abstract: Large-scale pretraining of transformers has been central to the success of foundation models.
By Ambroise Odonnat, Wassim Bouaziz, Vivien Cabannes
The paper demonstrates that applying sparsity-inducing regularization to multitask neural networks encourages the emergence of functional modularity without predefining module boundaries. By profiling multilayer perceptrons and a lightweight transformer, the authors identify compact, task‑selective subnetworks—called functional tokens—that can be selectively retained to preserve specific task performance while degrading others. These tokens tend to form in deeper layers, and their compactness and specialization are influenced by regularization strength, form, and task weighting, while they can be recombined to reconstruct composite computations without retraining.
By Bowen Shi, Qi Ouyang, Long Qian
arXiv:2609.38081v1 Announce Type: new
Abstract: On a single task, deep networks can learn many solutions, depending on their optimizer, training data, architecture, and hyperparameters. Many of these...
By Ann Huang, Mitchell Ostrow, Zhouyang Lu, William T. Redman, Leo Kozachkov, Kanaka Rajan
MoR-MLLM introduces a computation-sparse multimodal large language model that uses adaptive per-token recursion to dynamically adjust recursive depth. This approach allocates more computation to visually or linguistically challenging tokens while skipping redundant operations for simpler ones. A three-stage MoR-Tuning strategy and entropy-regularized loss stabilize training, enabling the model to reduce training memory and computation complexity while maintaining high performance on vision-language tasks.
By Pengcheng Zheng, Chaoning Zhang, Jiaxin Yan, Sihan Cao, Jianwei Zhang, Xudong Wang, Jiaquan Zhang, Jewon Lee, Tae-Ho Kim, Yang Yang, Heng Tao Shen
arXiv:2505.09716v4 Announce Type: replace
Abstract: Out-of-distribution (OOD) generalisation through composition requires a system to discover invariant properties from input-output associations and...
By George Dimitriadis, Spyridon Samothrakis
The paper investigates whether the sparsity of Mixture-of-Experts (MoE) models leads to intrinsic semantic organization across modalities and domains. It shows that experts naturally specialize semantically even without explicit modular training. The authors propose ExpertLens, a data‑free method that decodes router weights to identify domain‑specialized experts, enabling selective fine‑tuning that matches or exceeds full fine‑tuning while updating only 21.7–47.0% of parameters and achieving a 4.0× speedup, outperforming LoRA in both performance and efficiency.
By Damiano Marsili, Raphi Kang, Aditya Mehta, Pietro Perona, Georgia Gkioxari
arXiv:2607. 16295v1 Announce Type: cross Abstract: Mechanistic interpretability has made significant strides in understanding neural network representations, with sparse dictionary learning (SDL) methods, most prominently sparse autoencoders, as a central paradigm.
By Yiming Tang, Qinglin Qi, Zhaoqian Yao, Harshvardhan Saini, Dianbo Liu
Over the past decade, deep neural networks (DNNs) have achieved remarkable success on complex machine-learning tasks, yet the theoretical foundations of their performance remain incomplete. From a statistical viewpoint, a natural question is: can DNNs attain feature-learning and prediction consistency comparable to that of classical models?
arXiv:2604. 07753v2 Announce Type: replace-cross Abstract: Empowering Large Multimodal Models (LMMs) with image generation often leads to catastrophic forgetting in understanding tasks due to severe gradient conflicts.
By Xiangyue Liu, Zijian Zhang, Miles Yang, Zhao Zhong, Liefeng Bo, Ping Tan