Dynamical similarity analysis uniquely captures how computations develop in RNNs

• URL: http://arxiv.org/abs/2410.24070v2
• Date: Fri, 01 Nov 2024 09:41:09 GMT
• Title: Dynamical similarity analysis uniquely captures how computations develop in RNNs
• Authors: Quentin Guilhot, Michał Wójcik, Jascha Achterberg, Rui Ponte Costa,
• Abstract summary: Recent findings show that some metrics respond to spurious signals, leading to misleading results. We propose that compositional learning in recurrent neural networks (RNNs) can provide a test case for dynamical representation alignment metrics. We show that the recently proposed Dynamical Similarity Analysis (DSA) is more noise robust and reliably identifies behaviorally relevant representations.
• Score: 3.037387520023979
• License:
• Abstract: Methods for analyzing representations in neural systems are increasingly popular tools in neuroscience and mechanistic interpretability. Measures comparing neural activations across conditions, architectures, and species give scalable ways to understand information transformation within different neural networks. However, recent findings show that some metrics respond to spurious signals, leading to misleading results. Establishing benchmark test cases is thus essential for identifying the most reliable metric and potential improvements. We propose that compositional learning in recurrent neural networks (RNNs) can provide a test case for dynamical representation alignment metrics. Implementing this case allows us to evaluate if metrics can identify representations that develop throughout learning and determine if representations identified by metrics reflect the network's actual computations. Building both attractor and RNN based test cases, we show that the recently proposed Dynamical Similarity Analysis (DSA) is more noise robust and reliably identifies behaviorally relevant representations compared to prior metrics (Procrustes, CKA). We also demonstrate how such test cases can extend beyond metric evaluation to study new architectures. Specifically, testing DSA in modern (Mamba) state space models suggests that these models, unlike RNNs, may not require changes in recurrent dynamics due to their expressive hidden states. Overall, we develop test cases that showcase how DSA's enhanced ability to detect dynamical motifs makes it highly effective for identifying ongoing computations in RNNs and revealing how networks learn tasks.

Related papers

• Topological Representations of Heterogeneous Learning Dynamics of Recurrent Spiking Neural Networks [16.60622265961373]
  Spiking Neural Networks (SNNs) have become an essential paradigm in neuroscience and artificial intelligence. Recent advances in literature have studied the network representations of deep neural networks.
  arXiv  Detail & Related papers  (2024-03-19T05:37:26Z)
• Understanding Self-attention Mechanism via Dynamical System Perspective [58.024376086269015]
  Self-attention mechanism (SAM) is widely used in various fields of artificial intelligence. We show that intrinsic stiffness phenomenon (SP) in the high-precision solution of ordinary differential equations (ODEs) also widely exists in high-performance neural networks (NN) We show that the SAM is also a stiffness-aware step size adaptor that can enhance the model's representational ability to measure intrinsic SP.
  arXiv  Detail & Related papers  (2023-08-19T08:17:41Z)
• How neural networks learn to classify chaotic time series [77.34726150561087]
  We study the inner workings of neural networks trained to classify regular-versus-chaotic time series. We find that the relation between input periodicity and activation periodicity is key for the performance of LKCNN models.
  arXiv  Detail & Related papers  (2023-06-04T08:53:27Z)
• A Novel Explainable Out-of-Distribution Detection Approach for Spiking Neural Networks [6.100274095771616]
  This work presents a novel OoD detector that can identify whether test examples input to a Spiking Neural Network belong to the distribution of the data over which it was trained. We characterize the internal activations of the hidden layers of the network in the form of spike count patterns. A local explanation method is devised to produce attribution maps revealing which parts of the input instance push most towards the detection of an example as an OoD sample.
  arXiv  Detail & Related papers  (2022-09-30T11:16:35Z)
• Implicit N-grams Induced by Recurrence [10.053475465955794]
  We present a study that shows there actually exist some explainable components that reside within the hidden states. We evaluated such extracted explainable features from trained RNNs on downstream sentiment analysis tasks and found they could be used to model interesting linguistic phenomena.
  arXiv  Detail & Related papers  (2022-05-05T15:53:46Z)
• Data-driven emergence of convolutional structure in neural networks [83.4920717252233]
  We show how fully-connected neural networks solving a discrimination task can learn a convolutional structure directly from their inputs. By carefully designing data models, we show that the emergence of this pattern is triggered by the non-Gaussian, higher-order local structure of the inputs.
  arXiv  Detail & Related papers  (2022-02-01T17:11:13Z)
• On the benefits of robust models in modulation recognition [53.391095789289736]
  Deep Neural Networks (DNNs) using convolutional layers are state-of-the-art in many tasks in communications. In other domains, like image classification, DNNs have been shown to be vulnerable to adversarial perturbations. We propose a novel framework to test the robustness of current state-of-the-art models.
  arXiv  Detail & Related papers  (2021-03-27T19:58:06Z)
• Recoding latent sentence representations -- Dynamic gradient-based activation modification in RNNs [0.0]
  In RNNs, encoding information in a suboptimal way can impact the quality of representations based on later elements in the sequence. I propose an augmentation to standard RNNs in form of a gradient-based correction mechanism. I conduct different experiments in the context of language modeling, where the impact of using such a mechanism is examined in detail.
  arXiv  Detail & Related papers  (2021-01-03T17:54:17Z)
• The geometry of integration in text classification RNNs [20.76659136484842]
  We study recurrent networks trained on a battery of both natural and synthetic text classification tasks. We find the dynamics of these trained RNNs to be both interpretable and low-dimensional. Our observations span multiple architectures and datasets, reflecting a common mechanism RNNs employ to perform text classification.
  arXiv  Detail & Related papers  (2020-10-28T17:58:53Z)
• Explaining and Improving Model Behavior with k Nearest Neighbor Representations [107.24850861390196]
  We propose using k nearest neighbor representations to identify training examples responsible for a model's predictions. We show that kNN representations are effective at uncovering learned spurious associations. Our results indicate that the kNN approach makes the finetuned model more robust to adversarial inputs.
  arXiv  Detail & Related papers  (2020-10-18T16:55:25Z)
• Deep Representational Similarity Learning for analyzing neural signatures in task-based fMRI dataset [81.02949933048332]
  This paper develops Deep Representational Similarity Learning (DRSL), a deep extension of Representational Similarity Analysis (RSA) DRSL is appropriate for analyzing similarities between various cognitive tasks in fMRI datasets with a large number of subjects.
  arXiv  Detail & Related papers  (2020-09-28T18:30:14Z)

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.