Multiple Instance Learning (MIL) has been widely applied in pathology towards solving critical problems such as automating cancer diagnosis and grading, predicting patient prognosis, and therapy response. Deploying these models in a clinical setting requires careful inspection of these black boxes during development and deployment to identify failures and maintain physician trust. In this work, we propose a simple formulation of MIL models, which enables interpretability while maintaining similar predictive performance. Our Additive MIL models enable spatial credit assignment such that the contribution of each region in the image can be exactly computed and visualized. We show that our spatial credit assignment coincides with regions used by pathologists during diagnosis and improves upon classical attention heatmaps from attention MIL models. We show that any existing MIL model can be made additive with a simple change in function composition. We also show how these models can debug model failures, identify spurious features, and highlight class-wise regions of interest, enabling their use in high-stakes environments such as clinical decision-making.

A well-trained Convolutional Neural Network can easily be pruned without significant loss of performance. This is because of unnecessary overlap in the features captured by the network's filters. Innovations in network architecture such as skip/dense connections and inception units have mitigated this problem to some extent, but these improvements come with increased computation and memory requirements at run-time. We attempt to address this problem from another angle - not by changing the network structure but by altering the training method. We show that by temporarily pruning and then restoring a subset of the model's filters, and repeating this process cyclically, overlap in the learned features is reduced, producing improved generalization. We show that the existing model-pruning criteria are not optimal for selecting filters to prune in this context, and introduce inter-filter orthogonality as the ranking criteria to determine under-expressive filters. Our method is applicable both to vanilla convolutional networks and more complex modern architectures, and improves the performance across a variety of tasks, especially when applied to smaller networks.

We present a method to extend the idea of sprite coding to videos containing a wide variety of naturally occurring background motion, which could potentially be incorporated into existing and future video standards. The existing MPEG-4 part 2 standard, now almost 20 years old, provides the ability to store objects in separate layers, and includes a sprite mode where the background layer is generated by cropping a still image based on frame-wide global motion parameters, but videos containing more general background motion cannot be effectively encoded with sprite mode. We propose a perceptually motivated lossy compression algorithm, where oscillatory background motion can be compactly encoded. Our model achieves a low bit rate by referencing a time-invariant representation of the optical flow with only a few added parameters per frame. At very low bit rates, our technique can provide dynamic backgrounds at a visual quality that may not be achievable by traditional methods which are known to produce unacceptable blocking and ringing artifacts.

CNNs are poised to become integral parts of many critical systems. Despite their robustness to natural variations, image pixel values can be manipulated, via small, carefully crafted, imperceptible perturbations, to cause a model to misclassify images. We present an algorithm to process an image so that classification accuracy is significantly preserved in the presence of such adversarial manipulations. Image classifiers tend to be robust to natural noise, and adversarial attacks tend to be agnostic to object location. These observations motivate our strategy, which leverages model robustness to defend against adversarial perturbations by forcing the image to match natural image statistics. Our algorithm locally corrupts the image by redistributing pixel values via a process we term pixel deflection. A subsequent wavelet-based denoising operation softens this corruption, as well as some of the adversarial changes. We demonstrate experimentally that the combination of these techniques enables the effective recovery of the true class, against a variety of robust attacks. Our results compare favorably with current state-of-the-art defenses, without requiring retraining or modifying the CNN.

As deep neural networks (DNNs) have been integrated into critical systems, several methods to attack these systems have been developed. These adversarial attacks make imperceptible modifications to an image that fool DNN classifiers. We present an adaptive JPEG encoder which defends against many of these attacks. Experimentally, we show that our method produces images with high visual quality while greatly reducing the potency of state-of- the-art attacks. Our algorithm requires only a modest increase in encoding time, produces a compressed image which can be decompressed by an off-the-shelf JPEG decoder, and classified by an unmodified classifier.

Reza Ghaeini, Sadid A. Hasan, Vivek Datla, Joey Liu, Kathy Lee, Ashequl Qadir, Yuan Ling, Aaditya Prakash, Xiaoli Fern, Oladimeji Farri. Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers). 2018.

Diagnosis of a clinical condition is a challenging task, which often requires significant medical investigation. Previous work related to diagnostic inferencing problems mostly consider multivariate observational data (e.g. physiological signals, lab tests etc.). In contrast, we explore the problem using free-text medical notes recorded in an electronic health record (EHR). Complex tasks like these can benefit from structured knowledge bases, but those are not scalable. We instead exploit raw text from Wikipedia as a knowledge source. Memory networks have been demonstrated to be effective in tasks which require comprehension of free-form text. They use the final iteration of the learned representation to predict probable classes. We introduce condensed memory neural networks (C-MemNNs), a novel model with iterative condensation of memory representations that preserves the hierarchy of features in the memory. Experiments on the MIMIC-III dataset show that the proposed model outperforms other variants of memory networks to predict the most probable diagnoses given a complex clinical scenario.

It has long been considered a significant problem to improve the visual quality of lossy image and video compression. Recent advances in computing power together with the availability of large training data sets has increased interest in the application of deep learning CNNs to address image recognition and image processing tasks. Here, we present a powerful CNN tailored to the specific task of semantic image understanding to achieve higher visual quality in lossy compression. A modest increase in complexity is incorporated to the encoder which allows a standard, off-the-shelf JPEG decoder to be used. While JPEG encoding may be optimized for generic images, the process is ultimately unaware of the specific content of the image to be compressed. Our technique makes JPEG content-aware by designing and training a model to identify multiple semantic regions in a given image. Unlike object detection techniques, our model does not require labeling of object positions and is able to identify objects in a single pass. We present a new CNN architecture directed specifically to image compression, which generates a map that highlights semantically-salient regions so that they can be encoded at higher quality as compared to background regions. By adding a complete set of features for every class, and then taking a threshold over the sum of all feature activations, we generate a map that highlights semantically-salient regions so that they can be encoded at a better quality compared to background regions. Experiments are presented on the Kodak PhotoCD dataset and the MIT Saliency Benchmark dataset, in which our algorithm achieves higher visual quality for the same compressed size while preserving PSNR.

We present a new tool to help students make use of video recordings of traditional chalkboard and whiteboard lectures. Our tool provides a compact summary of a chalkboard lecture in the form of a set of slides, and an interactive environment for electronic note-taking in real time while watching a lecture in class or during review. Key to the processing of a chalkboard video by our system is the removal of the lecturer from the video (now no longer present to occlude the board) and segmentation of the resulting video to produce a set of slides comprising a compact presentation of the content written on the board during the lecture. Our method enhances learning in a number of ways. Students can focus on taking notes to supplement the material, rather than rote transcription. In addition, notes are aligned temporally with the lecture video and thus provide a useful way for students to index into the important parts of the video. Our system can be easily integrated into courses with a recorded lecture video component to help students organize their notes and quickly navigate the lectures from a given course for particular topics.

Current Adverse Drug Events (ADE) surveillance systems are often associated with a sizable time lag before such events are published. Online social media such as Twitter could describe adverse drug events in real-time, prior to official reporting. Deep learning has significantly improved text classification performance in recent years and can potentially enhance ADE classification in tweets. However, these models typically require large corpora with human expert-derived labels, and such resources are very expensive to generate and are hardly available. Semi-supervised deep learning models, which offer a plausible alternative to fully supervised models, involve the use of a small set of labeled data and a relatively larger collection of unlabeled data for training. Traditionally, these models are trained on labeled and unlabeled data from similar topics or domains. In reality, millions of tweets generated daily often focus on disparate topics, and this could present a challenge for building deep learning models for ADE classification with random Twitter stream as unlabeled training data. In this work, we build several semi-supervised convolutional neural network (CNN) models for ADE classification in tweets, specifically leveraging different types of unlabeled data in developing the models to address the problem. We demonstrate that, with the selective use of a variety of unlabeled data, our semi-supervised CNN models outperform a strong state-of-the-art supervised classification model by +9.9% F1-score. We evaluated our models on the Twitter data set used in the PSB 2016 Social Media Shared Task. Our results present the new state-of-the-art for this data set.

In this paper, we propose a novel neural approach for paraphrase generation. Conventional paraphrase generation methods either leverage hand-written rules and thesauri-based alignments, or use statistical machine learning principles. To the best of our knowledge, this work is the first to explore deep learning models for paraphrase generation. Our primary contribution is a stacked residual LSTM network, where we add residual connections between LSTM layers. This allows for efficient training of deep LSTMs. We evaluate our model and other state-of-the-art deep learning models on three different datasets: PPDB, WikiAnswers, and MSCOCO. Evaluation results demonstrate that our model outperforms sequence to sequence, attention-based, and bi-directional LSTM models on BLEU, METEOR, TER, and an embedding-based sentence similarity metric.