Glaucoma is the second leading cause of blindness in the world, impacting approximately 2.7 million people in the U.S alone. It is a complex set of diseases and, if left untreated, can lead to blindness. It's a particularly large issue in Australia, where only 50 percent of all people who have it are actually diagnosed and receive the treatment they need. As part of a team of scientists from IBM and New York University, my colleagues and I are looking at new ways AI could be used to help ophthalmologists and optometrists further utilize eye images, and potentially help to speed the process for detecting glaucoma in images. In a recent paper, we detail a new deep learning framework that detects glaucoma directly from raw optical coherence tomographic (OCT) imaging, a method which uses light waves to take cross-section pictures of the retina.

Recent studies have shown that slight perturbations in the input data can significantly affect the robustness of Deep Neural Networks (DNNs), leading to misclassification and confidence reduction. In this paper, we introduce a novel technique based on the Selective Sobel Convolution (SSC) operation in the training loop, that increases the robustness of a given DNN by allowing it to learn important edges in the input in a controlled fashion. This is achieved by introducing a trainable parameter, which acts as a threshold for eliminating the weaker edges. We validate our technique against the attacks of Cleverhans library on Convolutional DNNs against adversarial attacks. Our experimental results on the MNIST and CIFAR10 datasets illustrate that this controlled learning considerably increases the accuracy of the DNNs by 1.53% even when subjected to adversarial attacks.

Suppressing bones on chest X-rays such as ribs and clavicle is often expected to improve pathologies classification. These bones can interfere with a broad range of diagnostic tasks on pulmonary disease except for musculoskeletal system. Current conventional method for acquisition of bone suppressed X-rays is dual energy imaging, which captures two radiographs at a very short interval with different energy levels; however, the patient is exposed to radiation twice and the artifacts arise due to heartbeats between two shots. In this paper, we introduce a deep generative model trained to predict bone suppressed images on single energy chest X-rays, analyzing a finite set of previously acquired dual energy chest X-rays. Since the relatively small amount of data is available, such approach relies on the methodology maximizing the data utilization. Here we integrate the following two approaches. First, we use a conditional generative adversarial network that complements the traditional regression method minimizing the pairwise image difference. Second, we use Haar 2D wavelet decomposition to offer a perceptual guideline in frequency details to allow the model to converge quickly and efficiently. As a result, we achieve state-of-the-art performance on bone suppression as compared to the existing approaches with dual energy chest X-rays.

Amino acid sequence portrays most intrinsic form of a protein and expresses primary structure of protein. The order of amino acids in a sequence enables a protein to acquire a particular stable conformation that is responsible for the functions of the protein. This relationship between a sequence and its function motivates the need to analyse the sequences for predicting protein functions. Early generation computational methods using BLAST, FASTA, etc. perform function transfer based on sequence similarity with existing databases and are computationally slow. Although machine learning based approaches are fast, they fail to perform well for long protein sequences (i.e., protein sequences with more than 300 amino acid residues). In this paper, we introduce a novel method for construction of two separate feature sets for protein sequences based on analysis of 1) single fixed-sized segments and 2) multi-sized segments, using bi-directional long short-term memory network. Further, model based on proposed feature set is combined with the state of the art Multi-lable Linear Discriminant Analysis (MLDA) features based model to improve the accuracy. Extensive evaluations using separate datasets for biological processes and molecular functions demonstrate promising results for both single-sized and multi-sized segments based feature sets. While former showed an improvement of +3.37% and +5.48%, the latter produces an improvement of +5.38% and +8.00% respectively for two datasets over the state of the art MLDA based classifier. After combining two models, there is a significant improvement of +7.41% and +9.21% respectively for two datasets compared to MLDA based classifier. Specifically, the proposed approach performed well for the long protein sequences and superior overall performance.

Deep neural networks (DNNs) although achieving human-level performance in many domains, have very large model size that hinders their broader applications on edge computing devices. Extensive research work have been conducted on DNN model compression or pruning. However, most of the previous work took heuristic approaches. This work proposes a progressive weight pruning approach based on ADMM (Alternating Direction Method of Multipliers), a powerful technique to deal with non-convex optimization problems with potentially combinatorial constraints. Motivated by dynamic programming, the proposed method reaches extremely high pruning rate by using partial prunings with moderate pruning rates. Therefore, it resolves the accuracy degradation and long convergence time problems when pursuing extremely high pruning ratios. It achieves up to 34 times pruning rate for ImageNet dataset and 167 times pruning rate for MNIST dataset, significantly higher than those reached by the literature work. Under the same number of epochs, the proposed method also achieves faster convergence and higher compression rates. The codes and pruned DNN models are released in the link bit.ly/2zxdlss

The exponential increase in dependencies between the cyber and physical world leads to an enormous amount of data which must be efficiently processed and stored. Therefore, computing paradigms are evolving towards machine learning (ML)-based systems because of their ability to efficiently and accurately process the enormous amount of data. Although ML-based solutions address the efficient computing requirements of big data, they introduce (new) security vulnerabilities into the systems, which cannot be addressed by traditional monitoring-based security measures. Therefore, this paper first presents a brief overview of various security threats in machine learning, their respective threat models and associated research challenges to develop robust security measures. To illustrate the security vulnerabilities of ML during training, inferencing and hardware implementation, we demonstrate some key security threats on ML using LeNet and VGGNet for MNIST and German Traffic Sign Recognition Benchmarks (GTSRB), respectively. Moreover, based on the security analysis of ML-training, we also propose an attack that has a very less impact on the inference accuracy. Towards the end, we highlight the associated research challenges in developing security measures and provide a brief overview of the techniques used to mitigate such security threats.

While deep neural networks have achieved groundbreaking prediction results in many tasks, there is a class of data where existing architectures are not optimal -- sequences of probability distributions. Performing forward prediction on sequences of distributions has many important applications. However, there are two main challenges in designing a network model for this task. First, neural networks are unable to encode distributions compactly as each node encodes just a real value. A recent work of Distribution Regression Network (DRN) solved this problem with a novel network that encodes an entire distribution in a single node, resulting in improved accuracies while using much fewer parameters than neural networks. However, despite its compact distribution representation, DRN does not address the second challenge, which is the need to model time dependencies in a sequence of distributions. In this paper, we propose our Recurrent Distribution Regression Network (RDRN) which adopts a recurrent architecture for DRN. The combination of compact distribution representation and shared weights architecture across time steps makes RDRN suitable for modeling the time dependencies in a distribution sequence. Compared to neural networks and DRN, RDRN achieves the best prediction performance while keeping the network compact.

A basic problem in machine learning is to find a mapping $f$ from a low dimensional latent space to a high dimensional observation space. Equipped with the representation power of non-linearity, a learner can easily find a mapping which perfectly fits all the observations. However such a mapping is often not considered as good as it is not simple enough and over-fits. How to define simplicity? This paper tries to make such a formal definition of the amount of information imposed by a non-linear mapping. This definition is based on information geometry and is independent of observations, nor specific parametrizations. We prove these basic properties and discuss relationships with parametric and non-parametric embeddings.

Deep metric learning aims to learn a deep embedding that can capture the semantic similarity of data points. Given the availability of massive training samples, deep metric learning is known to suffer from slow convergence due to a large fraction of trivial samples. Therefore, most existing methods generally resort to sample mining strategies for selecting nontrivial samples to accelerate convergence and improve performance. In this work, we identify two critical limitations of the sample mining methods, and provide solutions for both of them. First, previous mining methods assign one binary score to each sample, i.e., dropping or keeping it, so they only selects a subset of relevant samples in a mini-batch. Therefore, we propose a novel sample mining method, called Online Soft Mining (OSM), which assigns one continuous score to each sample to make use of all samples in the mini-batch. OSM learns extended manifolds that preserve useful intraclass variances by focusing on more similar positives. Second, the existing methods are easily influenced by outliers as they are generally included in the mined subset. To address this, we introduce Class-Aware Attention (CAA) that assigns little attention to abnormal data samples. Furthermore, by combining OSM and CAA, we propose a novel weighted contrastive loss to learn discriminative embeddings. Extensive experiments on two fine-grained visual categorisation datasets and two video-based person re-identification benchmarks show that our method significantly outperforms the state-of-the-art.

Deep neural networks (DNN)-based machine learning (ML) algorithms have recently emerged as the leading ML paradigm particularly for the task of classification due to their superior capability of learning efficiently from large datasets. The discovery of a number of well-known attacks such as dataset poisoning, adversarial examples, and network manipulation (through the addition of malicious nodes) has, however, put the spotlight squarely on the lack of security in DNN-based ML systems. In particular, malicious actors can use these well-known attacks to cause random/targeted misclassification, or cause a change in the prediction confidence, by only slightly but systematically manipulating the environmental parameters, inference data, or the data acquisition block. Most of the prior adversarial attacks have, however, not accounted for the pre-processing noise filters commonly integrated with the ML-inference module. Our contribution in this work is to show that this is a major omission since these noise filters can render ineffective the majority of the existing attacks, which rely essentially on introducing adversarial noise. Apart from this, we also extend the state of the art by proposing a novel pre-processing noise Filter-aware Adversarial ML attack called FAdeML. To demonstrate the effectiveness of the proposed methodology, we generate an adversarial attack image by exploiting the "VGGNet" DNN trained for the "German Traffic Sign Recognition Benchmarks (GTSRB" dataset, which despite having no visual noise, can cause a classifier to misclassify even in the presence of pre-processing noise filters.