Mechanistic Design and Scaling of Hybrid Architectures

• URL: http://arxiv.org/abs/2403.17844v1
• Date: Tue, 26 Mar 2024 16:33:12 GMT
• Title: Mechanistic Design and Scaling of Hybrid Architectures
• Authors: Michael Poli, Armin W Thomas, Eric Nguyen, Pragaash Ponnusamy, Björn Deiseroth, Kristian Kersting, Taiji Suzuki, Brian Hie, Stefano Ermon, Christopher Ré, Ce Zhang, Stefano Massaroli,
• Abstract summary: We identify and test new hybrid architectures constructed from a variety of computational primitives. We experimentally validate the resulting architectures via an extensive compute-optimal and a new state-optimal scaling law analysis. We find MAD synthetics to correlate with compute-optimal perplexity, enabling accurate evaluation of new architectures.
• Score: 114.3129802943915
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The development of deep learning architectures is a resource-demanding process, due to a vast design space, long prototyping times, and high compute costs associated with at-scale model training and evaluation. We set out to simplify this process by grounding it in an end-to-end mechanistic architecture design (MAD) pipeline, encompassing small-scale capability unit tests predictive of scaling laws. Through a suite of synthetic token manipulation tasks such as compression and recall, designed to probe capabilities, we identify and test new hybrid architectures constructed from a variety of computational primitives. We experimentally validate the resulting architectures via an extensive compute-optimal and a new state-optimal scaling law analysis, training over 500 language models between 70M to 7B parameters. Surprisingly, we find MAD synthetics to correlate with compute-optimal perplexity, enabling accurate evaluation of new architectures via isolated proxy tasks. The new architectures found via MAD, based on simple ideas such as hybridization and sparsity, outperform state-of-the-art Transformer, convolutional, and recurrent architectures (Transformer++, Hyena, Mamba) in scaling, both at compute-optimal budgets and in overtrained regimes. Overall, these results provide evidence that performance on curated synthetic tasks can be predictive of scaling laws, and that an optimal architecture should leverage specialized layers via a hybrid topology.

Related papers

• AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation [48.82264764771652]
  We introduce AsCAN -- a hybrid architecture, combining both convolutional and transformer blocks. AsCAN supports a variety of tasks: recognition, segmentation, class-conditional image generation. We then scale the same architecture to solve a large-scale text-to-image task and show state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-11-07T18:43:17Z)
• A Realistic Simulation Framework for Analog/Digital Neuromorphic Architectures [73.65190161312555]
  ARCANA is a spiking neural network simulator designed to account for the properties of mixed-signal neuromorphic circuits. We show how the results obtained provide a reliable estimate of the behavior of the spiking neural network trained in software.
  arXiv  Detail & Related papers  (2024-09-23T11:16:46Z)
• Neural Architecture Codesign for Fast Bragg Peak Analysis [1.7081438846690533]
  We develop an automated pipeline to streamline neural architecture codesign for fast, real-time Bragg peak analysis in microscopy. Our method employs neural architecture search and AutoML to enhance these models, including hardware costs, leading to the discovery of more hardware-efficient neural architectures.
  arXiv  Detail & Related papers  (2023-12-10T19:42:18Z)
• Model-to-Circuit Cross-Approximation For Printed Machine Learning Classifiers [4.865819809855699]
  Printed electronics (PE) promises on-demand fabrication, low non-recurring engineering costs, and sub-cent fabrication costs. Large feature sizes in PE prohibit the realization of complex ML models in PE, even with bespoke architectures. We present an automated, cross-layer approximation framework tailored to bespoke architectures that enable complex ML models in PE.
  arXiv  Detail & Related papers  (2023-03-14T22:11:34Z)
• FlowNAS: Neural Architecture Search for Optical Flow Estimation [65.44079917247369]
  We propose a neural architecture search method named FlowNAS to automatically find the better encoder architecture for flow estimation task. Experimental results show that the discovered architecture with the weights inherited from the super-network achieves 4.67% F1-all error on KITTI.
  arXiv  Detail & Related papers  (2022-07-04T09:05:25Z)
• Hysteretic Behavior Simulation Based on Pyramid Neural Network:Principle, Network Architecture, Case Study and Explanation [0.0]
  A surrogate model based on neural networks shows significant potential in balancing efficiency and accuracy. Its serial information flow and prediction based on single-level features adversely affect the network performance. A weighted stacked pyramid neural network architecture is proposed herein.
  arXiv  Detail & Related papers  (2022-04-29T16:42:00Z)
• The Nonlinearity Coefficient -- A Practical Guide to Neural Architecture Design [3.04585143845864]
  We develop methods that can predict, without any training, whether an architecture will achieve a relatively high test or training error on a task after training. We then go on to explain the error in terms of the architecture definition itself and develop tools for modifying the architecture. Our first major contribution is to show that the 'degree of nonlinearity' of a neural architecture is a key causal driver behind its performance.
  arXiv  Detail & Related papers  (2021-05-25T20:47:43Z)
• STONNE: A Detailed Architectural Simulator for Flexible Neural Network Accelerators [5.326345912766044]
  STONNE is a cycle-accurate, highly-modular and highly-extensible simulation framework. We show how it can closely approach the performance results of the publicly available BSV-coded MAERI implementation.
  arXiv  Detail & Related papers  (2020-06-10T19:20:52Z)
• A Semi-Supervised Assessor of Neural Architectures [157.76189339451565]
  We employ an auto-encoder to discover meaningful representations of neural architectures. A graph convolutional neural network is introduced to predict the performance of architectures.
  arXiv  Detail & Related papers  (2020-05-14T09:02:33Z)
• Stage-Wise Neural Architecture Search [65.03109178056937]
  Modern convolutional networks such as ResNet and NASNet have achieved state-of-the-art results in many computer vision applications. These networks consist of stages, which are sets of layers that operate on representations in the same resolution. It has been demonstrated that increasing the number of layers in each stage improves the prediction ability of the network. However, the resulting architecture becomes computationally expensive in terms of floating point operations, memory requirements and inference time.
  arXiv  Detail & Related papers  (2020-04-23T14:16:39Z)

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.