Amirali Abdolrashidi

AmirAli Abdolrashidi, Lisa Wang, Shivani Agrawal et al.

Quantization has become a popular technique to compress neural networks and reduce compute cost, but most prior work focuses on studying quantization without changing the network size. Many real-world applications of neural networks have compute cost and memory budgets, which can be traded off with model quality by changing the number of parameters. In this work, we use ResNet as a case study to systematically investigate the effects of quantization on inference compute cost-quality tradeoff curves. Our results suggest that for each bfloat16 ResNet model, there are quantized models with lower cost and higher accuracy; in other words, the bfloat16 compute cost-quality tradeoff curve is Pareto-dominated by the 4-bit and 8-bit curves, with models primarily quantized to 4-bit yielding the best Pareto curve. Furthermore, we achieve state-of-the-art results on ImageNet for 4-bit ResNet-50 with quantization-aware training, obtaining a top-1 eval accuracy of 77.09%. We demonstrate the regularizing effect of quantization by measuring the generalization gap. The quantization method we used is optimized for practicality: It requires little tuning and is designed with hardware capabilities in mind. Our work motivates further research into optimal numeric formats for quantization, as well as the development of machine learning accelerators supporting these formats. As part of this work, we contribute a quantization library written in JAX, which is open-sourced at https://github.com/google-research/google-research/tree/master/aqt.

Yanqi Zhou, Sudip Roy, Amirali Abdolrashidi et al.

Most compilers for machine learning (ML) frameworks need to solve many correlated optimization problems to generate efficient machine code. Current ML compilers rely on heuristics based algorithms to solve these optimization problems one at a time. However, this approach is not only hard to maintain but often leads to sub-optimal solutions especially for newer model architectures. Existing learning based approaches in the literature are sample inefficient, tackle a single optimization problem, and do not generalize to unseen graphs making them infeasible to be deployed in practice. To address these limitations, we propose an end-to-end, transferable deep reinforcement learning method for computational graph optimization (GO), based on a scalable sequential attention mechanism over an inductive graph neural network. GO generates decisions on the entire graph rather than on each individual node autoregressively, drastically speeding up the search compared to prior methods. Moreover, we propose recurrent attention layers to jointly optimize dependent graph optimization tasks and demonstrate 33%-60% speedup on three graph optimization tasks compared to TensorFlow default optimization. On a diverse set of representative graphs consisting of up to 80,000 nodes, including Inception-v3, Transformer-XL, and WaveNet, GO achieves on average 21% improvement over human experts and 18% improvement over the prior state of the art with 15x faster convergence, on a device placement task evaluated in real systems.

Amirali Abdolrashidi, Mehdi Minaei, Elham Azimi et al.

Automatic prediction of age and gender from face images has drawn a lot of attention recently, due it is wide applications in various facial analysis problems. However, due to the large intra-class variation of face images (such as variation in lighting, pose, scale, occlusion), the existing models are still behind the desired accuracy level, which is necessary for the use of these models in real-world applications. In this work, we propose a deep learning framework, based on the ensemble of attentional and residual convolutional networks, to predict gender and age group of facial images with high accuracy rate. Using attention mechanism enables our model to focus on the important and informative parts of the face, which can help it to make a more accurate prediction. We train our model in a multi-task learning fashion, and augment the feature embedding of the age classifier, with the predicted gender, and show that doing so can further increase the accuracy of age prediction. Our model is trained on a popular face age and gender dataset, and achieved promising results. Through visualization of the attention maps of the train model, we show that our model has learned to become sensitive to the right regions of the face.

Shervin Minaee, Mehdi Minaei, Amirali Abdolrashidi

Generating realistic palmprint (more generally biometric) images has always been an interesting and, at the same time, challenging problem. Classical statistical models fail to generate realistic-looking palmprint images, as they are not powerful enough to capture the complicated texture representation of palmprint images. In this work, we present a deep learning framework based on generative adversarial networks (GAN), which is able to generate realistic palmprint images. To help the model learn more realistic images, we proposed to add a suitable regularization to the loss function, which imposes the line connectivity of generated palmprint images. This is very desirable for palmprints, as the principal lines in palm are usually connected. We apply this framework to a popular palmprint databases, and generate images which look very realistic, and similar to the samples in this database. Through experimental results, we show that the generated palmprint images look very realistic, have a good diversity, and are able to capture different parts of the prior distribution. We also report the Frechet Inception distance (FID) of the proposed model, and show that our model is able to achieve really good quantitative performance in terms of FID score.

Shervin Minaee, Amirali Abdolrashidi, Hang Su et al.

Deep learning-based models have been very successful in achieving state-of-the-art results in many of the computer vision, speech recognition, and natural language processing tasks in the last few years. These models seem a natural fit for handling the ever-increasing scale of biometric recognition problems, from cellphone authentication to airport security systems. Deep learning-based models have increasingly been leveraged to improve the accuracy of different biometric recognition systems in recent years. In this work, we provide a comprehensive survey of more than 120 promising works on biometric recognition (including face, fingerprint, iris, palmprint, ear, voice, signature, and gait recognition), which deploy deep learning models, and show their strengths and potentials in different applications. For each biometric, we first introduce the available datasets that are widely used in the literature and their characteristics. We will then talk about several promising deep learning works developed for that biometric, and show their performance on popular public benchmarks. We will also discuss some of the main challenges while using these models for biometric recognition, and possible future directions to which research in this area is headed.

Yanqi Zhou, Sudip Roy, Amirali Abdolrashidi et al.

Runtime and scalability of large neural networks can be significantly affected by the placement of operations in their dataflow graphs on suitable devices. With increasingly complex neural network architectures and heterogeneous device characteristics, finding a reasonable placement is extremely challenging even for domain experts. Most existing automated device placement approaches are impractical due to the significant amount of compute required and their inability to generalize to new, previously held-out graphs. To address both limitations, we propose an efficient end-to-end method based on a scalable sequential attention mechanism over a graph neural network that is transferable to new graphs. On a diverse set of representative deep learning models, including Inception-v3, AmoebaNet, Transformer-XL, and WaveNet, our method on average achieves 16% improvement over human experts and 9.2% improvement over the prior art with 15 times faster convergence. To further reduce the computation cost, we pre-train the policy network on a set of dataflow graphs and use a superposition network to fine-tune it on each individual graph, achieving state-of-the-art performance on large hold-out graphs with over 50k nodes, such as an 8-layer GNMT.

Shervin Minaee, Elham Azimi, Amirali Abdolrashidi

Fingerprint recognition has been utilized for cellphone authentication, airport security and beyond. Many different features and algorithms have been proposed to improve fingerprint recognition. In this paper, we propose an end-to-end deep learning framework for fingerprint recognition using convolutional neural networks (CNNs) which can jointly learn the feature representation and perform recognition. We train our model on a large-scale fingerprint recognition dataset, and improve over previous approaches in terms of accuracy. Our proposed model is able to achieve a very high recognition accuracy on a well-known fingerprint dataset. We believe this framework can be widely used for biometrics recognition tasks, making more scalable and accurate systems possible. We have also used a visualization technique to highlight the important areas in an input fingerprint image, that mostly impact the recognition results.

Shervin Minaee, Amirali Abdolrashidi

Iris recognition has been an active research area during last few decades, because of its wide applications in security, from airports to homeland security border control. Different features and algorithms have been proposed for iris recognition in the past. In this paper, we propose an end-to-end deep learning framework for iris recognition based on residual convolutional neural network (CNN), which can jointly learn the feature representation and perform recognition. We train our model on a well-known iris recognition dataset using only a few training images from each class, and show promising results and improvements over previous approaches. We also present a visualization technique which is able to detect the important areas in iris images which can mostly impact the recognition results. We believe this framework can be widely used for other biometrics recognition tasks, helping to have a more scalable and accurate systems.

Shervin Minaee, Elham Azimi, AmirAli Abdolrashidi

With the popularity of social networks, and e-commerce websites, sentiment analysis has become a more active area of research in the past few years. On a high level, sentiment analysis tries to understand the public opinion about a specific product or topic, or trends from reviews or tweets. Sentiment analysis plays an important role in better understanding customer/user opinion, and also extracting social/political trends. There has been a lot of previous works for sentiment analysis, some based on hand-engineering relevant textual features, and others based on different neural network architectures. In this work, we present a model based on an ensemble of long-short-term-memory (LSTM), and convolutional neural network (CNN), one to capture the temporal information of the data, and the other one to extract the local structure thereof. Through experimental results, we show that using this ensemble model we can outperform both individual models. We are also able to achieve a very high accuracy rate compared to the previous works.

Shervin Minaee, Amirali Abdolrashidi

Facial expression recognition has been an active research area over the past few decades, and it is still challenging due to the high intra-class variation. Traditional approaches for this problem rely on hand-crafted features such as SIFT, HOG and LBP, followed by a classifier trained on a database of images or videos. Most of these works perform reasonably well on datasets of images captured in a controlled condition, but fail to perform as good on more challenging datasets with more image variation and partial faces. In recent years, several works proposed an end-to-end framework for facial expression recognition, using deep learning models. Despite the better performance of these works, there still seems to be a great room for improvement. In this work, we propose a deep learning approach based on attentional convolutional network, which is able to focus on important parts of the face, and achieves significant improvement over previous models on multiple datasets, including FER-2013, CK+, FERG, and JAFFE. We also use a visualization technique which is able to find important face regions for detecting different emotions, based on the classifier's output. Through experimental results, we show that different emotions seems to be sensitive to different parts of the face.

Shervin Minaee, Amirali Abdolrashidi

Generating realistic biometric images has been an interesting and, at the same time, challenging problem. Classical statistical models fail to generate realistic-looking fingerprint images, as they are not powerful enough to capture the complicated texture representation in fingerprint images. In this work, we present a machine learning framework based on generative adversarial networks (GAN), which is able to generate fingerprint images sampled from a prior distribution (learned from a set of training images). We also add a suitable regularization term to the loss function, to impose the connectivity of generated fingerprint images. This is highly desirable for fingerprints, as the lines in each finger are usually connected. We apply this framework to two popular fingerprint databases, and generate images which look very realistic, and similar to the samples in those databases. Through experimental results, we show that the generated fingerprint images have a good diversity, and are able to capture different parts of the prior distribution. We also evaluate the Frechet Inception distance (FID) of our proposed model, and show that our model is able to achieve good quantitative performance in terms of this score.

Shervin Minaee, Amirali Abdolrashidi

Generating iris images which look realistic is both an interesting and challenging problem. Most of the classical statistical models are not powerful enough to capture the complicated texture representation in iris images, and therefore fail to generate iris images which look realistic. In this work, we present a machine learning framework based on generative adversarial network (GAN), which is able to generate iris images sampled from a prior distribution (learned from a set of training images). We apply this framework to two popular iris databases, and generate images which look very realistic, and similar to the image distribution in those databases. Through experimental results, we show that the generated iris images have a good diversity, and are able to capture different part of the prior distribution.

Shervin Minaee, Amirali Abdolrashidi, Yao Wang

Iris is one of the popular biometrics that is widely used for identity authentication. Different features have been used to perform iris recognition in the past. Most of them are based on hand-crafted features designed by biometrics experts. Due to tremendous success of deep learning in computer vision problems, there has been a lot of interest in applying features learned by convolutional neural networks on general image recognition to other tasks such as segmentation, face recognition, and object detection. In this paper, we have investigated the application of deep features extracted from VGG-Net for iris recognition. The proposed scheme has been tested on two well-known iris databases, and has shown promising results with the best accuracy rate of 99.4\%, which outperforms the previous best result.

Shervin Minaee, Amirali Abdolrashidi, Yao Wang

Face recognition has been an active research area in the past few decades. In general, face recognition can be very challenging due to variations in viewpoint, illumination, facial expression, etc. Therefore it is essential to extract features which are invariant to some or all of these variations. Here a new image representation, called scattering transform/network, has been used to extract features from faces. The scattering transform is a kind of convolutional network which provides a powerful multi-layer representation for signals. After extraction of scattering features, PCA is applied to reduce the dimensionality of the data and then a multi-class support vector machine is used to perform recognition. The proposed algorithm has been tested on three face datasets and achieved a very high recognition rate.

Shervin Minaee, Amirali Abdolrashidi, Yao Wang

Sparse decomposition has been extensively used for different applications including signal compression and denoising and document analysis. In this paper, sparse decomposition is used for image segmentation. The proposed algorithm separates the background and foreground using a sparse-smooth decomposition technique such that the smooth and sparse components correspond to the background and foreground respectively. This algorithm is tested on several test images from HEVC test sequences and is shown to have superior performance over other methods, such as the hierarchical k-means clustering in DjVu. This segmentation algorithm can also be used for text extraction, video compression and medical image segmentation.

Shervin Minaee, AmirAli Abdolrashidi, Yao Wang

Iris recognition has drawn a lot of attention since the mid-twentieth century. Among all biometric features, iris is known to possess a rich set of features. Different features have been used to perform iris recognition in the past. In this paper, two powerful sets of features are introduced to be used for iris recognition: scattering transform-based features and textural features. PCA is also applied on the extracted features to reduce the dimensionality of the feature vector while preserving most of the information of its initial value. Minimum distance classifier is used to perform template matching for each new test sample. The proposed scheme is tested on a well-known iris database, and showed promising results with the best accuracy rate of 99.2%.

Shervin Minaee, AmirAli Abdolrashidi

Biometric-based identification has drawn a lot of attention in the recent years. Among all biometrics, palmprint is known to possess a rich set of features. In this paper we have proposed to use DCT-based features in parallel with wavelet-based ones for palmprint identification. PCA is applied to the features to reduce their dimensionality and the majority voting algorithm is used to perform classification. The features introduced here result in a near-perfectly accurate identification. This method is tested on a well-known multispectral palmprint database and an accuracy rate of 99.97-100\% is achieved, outperforming all previous methods in similar conditions.

Shervin Minaee, AmirAli Abdolrashidi

In order to utilize identification to the best extent, we need robust and fast algorithms and systems to process the data. Having palmprint as a reliable and unique characteristic of every person, we extract and use its features based on its geometry, lines and angles. There are countless ways to define measures for the recognition task. To analyze a new point of view, we extracted textural features and used them for palmprint recognition. Co-occurrence matrix can be used for textural feature extraction. As classifiers, we have used the minimum distance classifier (MDC) and the weighted majority voting system (WMV). The proposed method is tested on a well-known multispectral palmprint dataset of 6000 samples and an accuracy rate of 99.96-100% is obtained for most scenarios which outperforms all previous works in multispectral palmprint recognition.

Shervin Minaee, AmirAli Abdolrashidi

Palmprint is one of the most useful physiological biometrics that can be used as a powerful means in personal recognition systems. The major features of the palmprints are palm lines, wrinkles and ridges, and many approaches use them in different ways towards solving the palmprint recognition problem. Here we have proposed to use a set of statistical and wavelet-based features; statistical to capture the general characteristics of palmprints; and wavelet-based to find those information not evident in the spatial domain. Also we use two different classification approaches, minimum distance classifier scheme and weighted majority voting algorithm, to perform palmprint matching. The proposed method is tested on a well-known palmprint dataset of 6000 samples and has shown an impressive accuracy rate of 99.65\%-100\% for most scenarios.