skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Recurrent networks recognize patterns with low-dimensional oscillations
This study proposes a novel dynamical mechanism for pattern recognition discovered by interpreting a recurrent neural network (RNN) trained on a simple task inspired by the SET card game. We interpreted the trained RNN as recognizing patterns via phase shifts in a low-dimensional limit cycle in a manner analogous to transitions in a finite state automaton (FSA). We further validated this interpretation by handcrafting a simple oscillatory model that reproduces the dynamics of the trained RNN. Our findings not only suggest of a potential dynamical mechanism capable of pattern recognition, but also suggest of a potential neural implementation of FSA. Above all, this work contributes to the growing discourse on deep learning model interpretability.  more » « less
Award ID(s):
2139936 2003830
PAR ID:
10501663
Author(s) / Creator(s):
Publisher / Repository:
IEEE 22nd International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC'23)
Date Published:
Journal Name:
IEEE 22nd International Conference on Cognitive Informatics and Cognitive Computing (ICCI*CC'23)
Format(s):
Medium: X
Location:
Stanford University, California
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; (, Advances in neural information processing systems)
    null (Ed.)
    Recurrent neural networks (RNNs) are a widely used tool for modeling sequential data, yet they are often treated as inscrutable black boxes. Given a trained recurrent network, we would like to reverse engineer it--to obtain a quantitative, interpretable description of how it solves a particular task. Even for simple tasks, a detailed understanding of how recurrent networks work, or a prescription for how to develop such an understanding, remains elusive. In this work, we use tools from dynamical systems analysis to reverse engineer recurrent networks trained to perform sentiment classification, a foundational natural language processing task. Given a trained network, we find fixed points of the recurrent dynamics and linearize the nonlinear system around these fixed points. Despite their theoretical capacity to implement complex, high-dimensional computations, we find that trained networks converge to highly interpretable, low-dimensional representations. In particular, the topological structure of the fixed points and corresponding linearized dynamics reveal an approximate line attractor within the RNN, which we can use to quantitatively understand how the RNN solves the sentiment analysis task. Finally, we find this mechanism present across RNN architectures (including LSTMs, GRUs, and vanilla RNNs) trained on multiple datasets, suggesting that our findings are not unique to a particular architecture or dataset. Overall, these results demonstrate that surprisingly universal and human interpretable computations can arise across a range of recurrent networks. 
    more » « less
  2. ; ; ; ; (, Advances in neural information processing systems)
    null (Ed.)
    Many recent studies have employed task-based modeling with recurrent neural networks (RNNs) to infer the computational function of different brain regions. These models are often assessed by quantitatively comparing the low-dimensional neural dynamics of the model and the brain, for example using canonical correlation analysis (CCA). However, the nature of the detailed neurobiological inferences one can draw from such efforts remains elusive. For example, to what extent does training neural networks to solve simple tasks, prevalent in neuroscientific studies, uniquely determine the low-dimensional dynamics independent of neural architectures? Or alternatively, are the learned dynamics highly sensitive to different neural architectures? Knowing the answer to these questions has strong implications on whether and how to use task-based RNN modeling to understand brain dynamics. To address these foundational questions, we study populations of thousands of networks of commonly used RNN architectures trained to solve neuroscientifically motivated tasks and characterize their low-dimensional dynamics via CCA and nonlinear dynamical systems analysis. We find the geometry of the dynamics can be highly sensitive to different network architectures, and further find striking dissociations between geometric similarity as measured by CCA and network function, yielding a cautionary tale. Moreover, we find that while the geometry of neural dynamics can vary greatly across architectures, the underlying computational scaffold: the topological structure of fixed points, transitions between them, limit cycles, and linearized dynamics, often appears {\it universal} across all architectures. Overall, this analysis of universality and individuality across large populations of RNNs provides a much needed foundation for interpreting quantitative measures of dynamical similarity between RNN and brain dynamics. 
    more » « less
  3. ; ; (, PLOS Computational Biology)
    Theunissen, Frédéric E (Ed.)
    Human speech recognition transforms a continuous acoustic signal into categorical linguistic units, by aggregating information that is distributed in time. It has been suggested that this kind of information processing may be understood through the computations of a Recurrent Neural Network (RNN) that receives input frame by frame, linearly in time, but builds an incremental representation of this input through a continually evolving internal state. While RNNs can simulate several keybehavioralobservations about human speech and language processing, it is unknown whether RNNs also develop computational dynamics that resemble humanneural speech processing. Here we show that the internal dynamics of long short-term memory (LSTM) RNNs, trained to recognize speech from auditory spectrograms, predict human neural population responses to the same stimuli, beyond predictions from auditory features. Variations in the RNN architecture motivated by cognitive principles further improved this predictive power. Specifically, modifications that allow more human-like phonetic competition also led to more human-like temporal dynamics. Overall, our results suggest that RNNs provide plausible computational models of the cortical processes supporting human speech recognition. 
    more » « less
  4. ; (, Entropy)
    Recurrent neural networks (RNNs) are nonlinear dynamical models commonly used in the machine learning and dynamical systems literature to represent complex dynamical or sequential relationships between variables. Recently, as deep learning models have become more common, RNNs have been used to forecast increasingly complicated systems. Dynamical spatio-temporal processes represent a class of complex systems that can potentially benefit from these types of models. Although the RNN literature is expansive and highly developed, uncertainty quantification is often ignored. Even when considered, the uncertainty is generally quantified without the use of a rigorous framework, such as a fully Bayesian setting. Here we attempt to quantify uncertainty in a more formal framework while maintaining the forecast accuracy that makes these models appealing, by presenting a Bayesian RNN model for nonlinear spatio-temporal forecasting. Additionally, we make simple modifications to the basic RNN to help accommodate the unique nature of nonlinear spatio-temporal data. The proposed model is applied to a Lorenz simulation and two real-world nonlinear spatio-temporal forecasting applications. 
    more » « less
  5. (, Proceedings of 2018 International Symposium on Flexible Automation)
    Fault pattern recognition in complex mechanical systems such as gearbox has always been a great challenge. The performance of a classic fault pattern recognition approach heavily depends on domain expertise and the classifier applied. This paper proposes a deep convolutional neural network-based transfer learning approach that not only entertains adaptive feature extractions, but also requires only a small set of training data. The proposed transfer learning architecture essentially consists of two sequentially connected pieces; first is of a pre-trained deep neural network that serves to extract features automatically, the second piece is a neural network aimed for classification which is to be trained using data collected from gearbox experiment. The proposed approach performs gear fault pattern recognition using raw accelerometer data. The achieved accuracy indicates that the proposed approach is not only sensitive and robust in performance, but also has the potential to be applied to other pattern recognition practices. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email