Pruning and quantization are core techniques used to reduce the inference costs of deep neural networks. Among the state-of-the-art pruning techniques, magnitude-based pruning algorithms have demonstrated consistent success in the reduction of both weight and feature map complexity. However, we find that existing measures of neuron (or channel) importance estimation used for such pruning procedures have at least one of two limitations: (1) failure to consider the interdependence between successive layers; and/or (2) performing the estimation in a parametric setting or by using distributional assumptions on the feature maps. In this work, we demonstrate that the importance rankings of the output neurons of a given layer strongly depend on the sparsity level of the preceding layer, and therefore, naïvely estimating neuron importance to drive magnitude-based pruning will lead to suboptimal performance. Informed by this observation, we propose a purely data-driven nonparametric, magnitude-based channel pruning strategy that works in a greedy manner based on the activations of the previous sparsified layer. We demonstrate that our proposed method works effectively in combination with statistics-based quantization techniques to generate low precision structured subnetworks that can be efficiently accelerated by hardware platforms such as GPUs and FPGAs. Using our proposed algorithms, we demonstrate increased performance per memory footprint over existing solutions across a range of discriminative and generative networks.
Restricted Boltzmann Machines and Deep Belief Networks have been successfully used in a wide variety of applications including image classification and speech recognition. Inference and learning in these algorithms uses a Markov Chain Monte Carlo procedure called Gibbs sampling. A sigmoidal function forms the kernel of this sampler which can be realized from the firing statistics of noisy integrate-and-fire neurons on a neuromorphic VLSI substrate. This paper demonstrates such an implementation on an array of digital spiking neurons with stochastic leak and threshold properties for inference tasks and presents some key performance metrics for such a hardware-based sampler in both the generative and discriminative contexts.
X-ray Computed Tomography (XCT) plays a vital role in characterizing internal porosity in laser Powder Bed Fusion (PBF-L) parts, where defects critically impact mechanical properties. Traditional methods for porosity segmentation rely on supervised machine learning, which requires large labeled datasets and high computational resources, limiting scalability. Foundation Models, pre-trained on diverse datasets, offer a flexible alternative by enabling high accuracy through user-provided prompts with minimal fine-tuning. This study develops a novel unsupervised porosity segmentation framework based on the Segment Anything Model (SAM), a Vision Transformer-based Foundation Model. Utilizing a multi-point prompt generation scheme with unsupervised clustering, it achieves a Dice Similarity Coefficient (DSC) of over 80%, enabling efficient defect segmentation without labeled data. Additionally, the framework’s performance is validated across varied prompt sets by bootstrapping for uncertainty quantification. By addressing scalability and automation challenges, this work highlights the transformative potential of Foundation Models in enhancing XCT-based porosity characterization for PBF-L parts.
State-of-the-art deep learning solutions for image upsampling are currently trained using either resize or sub-pixel convolution to learn kernels that generate high fidelity images with minimal artifacts. However, performing inference with these learned convolution kernels requires memory-intensive feature map transformations that dominate time and energy costs in real-time applications. To alleviate this pressure on memory bandwidth, we propose a novel energy-efficient edge computing paradigm that confines the use of resize or sub-pixel convolution to training in the cloud by transforming learned convolution kernels to deconvolution kernels before deploying them for inference as a functionally equivalent deconvolution. These kernel transformations, intended as a one-time cost when shifting from training to inference, enable a systems designer to use each algorithm in their optimal context by preserving the image fidelity learned when training in the cloud while minimizing data transfer penalties during inference at the edge. We compare the inference properties of these convolution-based image upsampling algorithms and introduce a novel deconvolution inference algorithm, which we refer to as REVD2. To demonstrate the benefits of our approach, we upsample images selected from the BSD300 dataset using a pre-trained single-image super resolution network provided by the PyTorch model zoo. Using quantitative models of incurred time and energy costs to analyze this deep neural network, we estimate that using REVD2 for inference at the edge improves system latency by 2.1x or 2.8x and energy efficiency by 2.1x or 2.7x when respectively compared to sub-pixel or resize convolution counterparts.