Models, code, and papers for "Audrey G":

##### Fully-Automatic Semantic Segmentation for Food Intake Tracking in Long-Term Care Homes

Malnutrition impacts quality of life and places annually-recurring burden on the health care system. Half of older adults are at risk for malnutrition in long-term care (LTC). Monitoring and measuring nutritional intake is paramount yet involves time-consuming and subjective visual assessment, limiting current methods' reliability. The opportunity for automatic image-based estimation exists. Some progress outside LTC has been made (e.g., calories consumed, food classification), however, these methods have not been implemented in LTC, potentially due to a lack of ability to independently evaluate automatic segmentation methods within the intake estimation pipeline. Here, we propose and evaluate a novel fully-automatic semantic segmentation method for pixel-level classification of food on a plate using a deep convolutional neural network (DCNN). The macroarchitecture of the DCNN is a multi-scale encoder-decoder food network (EDFN) architecture comprising a residual encoder microarchitecture, a pyramid scene parsing decoder microarchitecture, and a specialized per-pixel food/no-food classification layer. The network was trained and validated on the pre-labelled UNIMIB 2016 food dataset (1027 tray images, 73 categories), and tested on our novel LTC plate dataset (390 plate images, 9 categories). Our fully-automatic segmentation method attained similar intersection over union to the semi-automatic graph cuts (91.2% vs. 93.7%). Advantages of our proposed system include: testing on a novel dataset, decoupled error analysis, no user-initiated annotations, with similar segmentation accuracy and enhanced reliability in terms of types of segmentation errors. This may address several short-comings currently limiting utility of automated food intake tracking in time-constrained LTC and hospital settings.

##### Nature vs. Nurture: The Role of Environmental Resources in Evolutionary Deep Intelligence

Feb 09, 2018
Audrey G. Chung, Paul Fieguth, Alexander Wong

Evolutionary deep intelligence synthesizes highly efficient deep neural networks architectures over successive generations. Inspired by the nature versus nurture debate, we propose a study to examine the role of external factors on the network synthesis process by varying the availability of simulated environmental resources. Experimental results were obtained for networks synthesized via asexual evolutionary synthesis (1-parent) and sexual evolutionary synthesis (2-parent, 3-parent, and 5-parent) using a 10% subset of the MNIST dataset. Results show that a lower environmental factor model resulted in a more gradual loss in performance accuracy and decrease in storage size. This potentially allows significantly reduced storage size with minimal to no drop in performance accuracy, and the best networks were synthesized using the lowest environmental factor models.

##### EdgeSpeechNets: Highly Efficient Deep Neural Networks for Speech Recognition on the Edge

Oct 18, 2018
Zhong Qiu Lin, Audrey G. Chung, Alexander Wong

Despite showing state-of-the-art performance, deep learning for speech recognition remains challenging to deploy in on-device edge scenarios such as mobile and other consumer devices. Recently, there have been greater efforts in the design of small, low-footprint deep neural networks (DNNs) that are more appropriate for edge devices, with much of the focus on design principles for hand-crafting efficient network architectures. In this study, we explore a human-machine collaborative design strategy for building low-footprint DNN architectures for speech recognition through a marriage of human-driven principled network design prototyping and machine-driven design exploration. The efficacy of this design strategy is demonstrated through the design of a family of highly-efficient DNNs (nicknamed EdgeSpeechNets) for limited-vocabulary speech recognition. Experimental results using the Google Speech Commands dataset for limited-vocabulary speech recognition showed that EdgeSpeechNets have higher accuracies than state-of-the-art DNNs (with the best EdgeSpeechNet achieving ~97% accuracy), while achieving significantly smaller network sizes (as much as 7.8x smaller) and lower computational cost (as much as 36x fewer multiply-add operations, 10x lower prediction latency, and 16x smaller memory footprint on a Motorola Moto E phone), making them very well-suited for on-device edge voice interface applications.

* 4 pages
##### A new take on measuring relative nutritional density: The feasibility of using a deep neural network to assess commercially-prepared pureed food concentrations

Dysphagia affects 590 million people worldwide and increases risk for malnutrition. Pureed food may reduce choking, however preparation differences impact nutrient density making quality assurance necessary. This paper is the first study to investigate the feasibility of computational pureed food nutritional density analysis using an imaging system. Motivated by a theoretical optical dilution model, a novel deep neural network (DNN) was evaluated using 390 samples from thirteen types of commercially prepared purees at five dilutions. The DNN predicted relative concentration of the puree sample (20%, 40%, 60%, 80%, 100% initial concentration). Data were captured using same-side reflectance of multispectral imaging data at different polarizations at three exposures. Experimental results yielded an average top-1 prediction accuracy of 92.2+/-0.41% with sensitivity and specificity of 83.0+/-15.0% and 95.0+/-4.8%, respectively. This DNN imaging system for nutrient density analysis of pureed food shows promise as a novel tool for nutrient quality assurance.

* 26 pages
##### ProstateGAN: Mitigating Data Bias via Prostate Diffusion Imaging Synthesis with Generative Adversarial Networks

Generative Adversarial Networks (GANs) have shown considerable promise for mitigating the challenge of data scarcity when building machine learning-driven analysis algorithms. Specifically, a number of studies have shown that GAN-based image synthesis for data augmentation can aid in improving classification accuracy in a number of medical image analysis tasks, such as brain and liver image analysis. However, the efficacy of leveraging GANs for tackling prostate cancer analysis has not been previously explored. Motivated by this, in this study we introduce ProstateGAN, a GAN-based model for synthesizing realistic prostate diffusion imaging data. More specifically, in order to generate new diffusion imaging data corresponding to a particular cancer grade (Gleason score), we propose a conditional deep convolutional GAN architecture that takes Gleason scores into consideration during the training process. Experimental results show that high-quality synthetic prostate diffusion imaging data can be generated using the proposed ProstateGAN for specified Gleason scores.

* Machine Learning for Health (ML4H) Workshop at NeurIPS 2018 arXiv:1811.07216
##### Discovery Radiomics for Pathologically-Proven Computed Tomography Lung Cancer Prediction

Lung cancer is the leading cause for cancer related deaths. As such, there is an urgent need for a streamlined process that can allow radiologists to provide diagnosis with greater efficiency and accuracy. A powerful tool to do this is radiomics: a high-dimension imaging feature set. In this study, we take the idea of radiomics one step further by introducing the concept of discovery radiomics for lung cancer prediction using CT imaging data. In this study, we realize these custom radiomic sequencers as deep convolutional sequencers using a deep convolutional neural network learning architecture. To illustrate the prognostic power and effectiveness of the radiomic sequences produced by the discovered sequencer, we perform cancer prediction between malignant and benign lesions from 97 patients using the pathologically-proven diagnostic data from the LIDC-IDRI dataset. Using the clinically provided pathologically-proven data as ground truth, the proposed framework provided an average accuracy of 77.52% via 10-fold cross-validation with a sensitivity of 79.06% and specificity of 76.11%, surpassing the state-of-the art method.

* 8 pages
##### Discovery Radiomics via StochasticNet Sequencers for Cancer Detection

Radiomics has proven to be a powerful prognostic tool for cancer detection, and has previously been applied in lung, breast, prostate, and head-and-neck cancer studies with great success. However, these radiomics-driven methods rely on pre-defined, hand-crafted radiomic feature sets that can limit their ability to characterize unique cancer traits. In this study, we introduce a novel discovery radiomics framework where we directly discover custom radiomic features from the wealth of available medical imaging data. In particular, we leverage novel StochasticNet radiomic sequencers for extracting custom radiomic features tailored for characterizing unique cancer tissue phenotype. Using StochasticNet radiomic sequencers discovered using a wealth of lung CT data, we perform binary classification on 42,340 lung lesions obtained from the CT scans of 93 patients in the LIDC-IDRI dataset. Preliminary results show significant improvement over previous state-of-the-art methods, indicating the potential of the proposed discovery radiomics framework for improving cancer screening and diagnosis.

* 3 pages

* 8 pages
##### Development of spatial suppression surrounding the focus of visual attention

Sep 17, 2018
Audrey M. B. Wong-Kee-You, John K. Tsotsos, Scott A. Adler

The capacity to filter out irrelevant information from our environment is critical to efficient processing. Yet, during development, when building a knowledge base of the world is occurring, the ability to selectively allocate attentional resources is limited (e.g., Amso & Scerif, 2015). In adulthood, research has demonstrated that surrounding the spatial location of attentional focus is a suppressive field, resulting from top-down attention promoting the processing of relevant stimuli and inhibiting surrounding distractors (e.g., Hopf et al., 2006). It is not fully known, however, whether this phenomenon manifests in development. In the current study, we examined whether spatial suppression surrounding the focus of visual attention is exhibited in developmental age groups. Participants between 12 and 27 years of age exhibited spatial suppression surrounding their focus of visual attention. Their accuracy increased as a function of the separation distance between a spatially cued (and attended) target and a second target, suggesting that a ring of suppression surrounded the attended target. When a central cue was instead presented and therefore attention was no longer spatially cued, surround suppression was not observed, indicating that our initial findings of suppression were indeed related to the focus of attention. Attentional surround suppression was not observed in 8- to 11-years-olds, even with a longer spatial cue presentation time, demonstrating that the lack of the effect at these ages is not due to slowed attentional feedback processes. Our findings demonstrate that top-down attentional processes are still immature until approximately 12 years of age, and that they continue to be refined throughout adolescence, converging well with previous research on attentional development.

##### Leveraging exploration in off-policy algorithms via normalizing flows

May 16, 2019
Bogdan Mazoure, Thang Doan, Audrey Durand, R Devon Hjelm, Joelle Pineau

Exploration is a crucial component for discovering approximately optimal policies in most high-dimensional reinforcement learning (RL) settings with sparse rewards. Approaches such as neural density models and continuous exploration (e.g., Go-Explore) have been instrumental in recent advances. Soft actor-critic (SAC) is a method for improving exploration that aims to combine off-policy updates while maximizing the policy entropy. We extend SAC to a richer class of probability distributions through normalizing flows, which we show improves performance in exploration, sample complexity, and convergence. Finally, we show that not only the normalizing flow policy outperforms SAC on MuJoCo domains, it is also significantly lighter, using as low as 5.6% of the original network's parameters for similar performance.

##### Estimating Quality in Multi-Objective Bandits Optimization

Apr 20, 2017
Audrey Durand, Christian Gagné

Many real-world applications are characterized by a number of conflicting performance measures. As optimizing in a multi-objective setting leads to a set of non-dominated solutions, a preference function is required for selecting the solution with the appropriate trade-off between the objectives. The question is: how good do estimations of these objectives have to be in order for the solution maximizing the preference function to remain unchanged? In this paper, we introduce the concept of preference radius to characterize the robustness of the preference function and provide guidelines for controlling the quality of estimations in the multi-objective setting. More specifically, we provide a general formulation of multi-objective optimization under the bandits setting. We show how the preference radius relates to the optimal gap and we use this concept to provide a theoretical analysis of the Thompson sampling algorithm from multivariate normal priors. We finally present experiments to support the theoretical results and highlight the fact that one cannot simply scalarize multi-objective problems into single-objective problems.

* Submitted to ECML 2017
##### Assessing Architectural Similarity in Populations of Deep Neural Networks

Apr 19, 2019
Audrey Chung, Paul Fieguth, Alexander Wong

Evolutionary deep intelligence has recently shown great promise for producing small, powerful deep neural network models via the synthesis of increasingly efficient architectures over successive generations. Despite recent research showing the efficacy of multi-parent evolutionary synthesis, little has been done to directly assess architectural similarity between networks during the synthesis process for improved parent network selection. In this work, we present a preliminary study into quantifying architectural similarity via the percentage overlap of architectural clusters. Results show that networks synthesized using architectural alignment (via gene tagging) maintain higher architectural similarities within each generation, potentially restricting the search space of highly efficient network architectures.

* 3 pages. arXiv admin note: text overlap with arXiv:1811.07966
##### Mitigating Architectural Mismatch During the Evolutionary Synthesis of Deep Neural Networks

Nov 19, 2018
Audrey Chung, Paul Fieguth, Alexander Wong

Evolutionary deep intelligence has recently shown great promise for producing small, powerful deep neural network models via the organic synthesis of increasingly efficient architectures over successive generations. Existing evolutionary synthesis processes, however, have allowed the mating of parent networks independent of architectural alignment, resulting in a mismatch of network structures. We present a preliminary study into the effects of architectural alignment during evolutionary synthesis using a gene tagging system. Surprisingly, the network architectures synthesized using the gene tagging approach resulted in slower decreases in performance accuracy and storage size; however, the resultant networks were comparable in size and performance accuracy to the non-gene tagging networks. Furthermore, we speculate that there is a noticeable decrease in network variability for networks synthesized with gene tagging, indicating that enforcing a like-with-like mating policy potentially restricts the exploration of the search space of possible network architectures.

* 5 pages
##### Logics of Temporal-Epistemic Actions

Nov 23, 2014
Bryan Renne, Joshua Sack, Audrey Yap

We present Dynamic Epistemic Temporal Logic, a framework for reasoning about operations on multi-agent Kripke models that contain a designated temporal relation. These operations are natural extensions of the well-known "action models" from Dynamic Epistemic Logic. Our "temporal action models" may be used to define a number of informational actions that can modify the "objective" temporal structure of a model along with the agents' basic and higher-order knowledge and beliefs about this structure, including their beliefs about the time. In essence, this approach provides one way to extend the domain of action model-style operations from atemporal Kripke models to temporal Kripke models in a manner that allows actions to control the flow of time. We present a number of examples to illustrate the subtleties involved in interpreting the effects of our extended action models on temporal Kripke models. We also study preservation of important epistemic-temporal properties of temporal Kripke models under temporal action model-induced operations, provide complete axiomatizations for two theories of temporal action models, and connect our approach with previous work on time in Dynamic Epistemic Logic.

##### Streaming kernel regression with provably adaptive mean, variance, and regularization

Aug 02, 2017
Audrey Durand, Odalric-Ambrym Maillard, Joelle Pineau

We consider the problem of streaming kernel regression, when the observations arrive sequentially and the goal is to recover the underlying mean function, assumed to belong to an RKHS. The variance of the noise is not assumed to be known. In this context, we tackle the problem of tuning the regularization parameter adaptively at each time step, while maintaining tight confidence bounds estimates on the value of the mean function at each point. To this end, we first generalize existing results for finite-dimensional linear regression with fixed regularization and known variance to the kernel setup with a regularization parameter allowed to be a measurable function of past observations. Then, using appropriate self-normalized inequalities we build upper and lower bound estimates for the variance, leading to Bersntein-like concentration bounds. The later is used in order to define the adaptive regularization. The bounds resulting from our technique are valid uniformly over all observation points and all time steps, and are compared against the literature with numerical experiments. Finally, the potential of these tools is illustrated by an application to kernelized bandits, where we revisit the Kernel UCB and Kernel Thompson Sampling procedures, and show the benefits of the novel adaptive kernel tuning strategy.

##### Old Dog Learns New Tricks: Randomized UCB for Bandit Problems

Oct 11, 2019
Sharan Vaswani, Abbas Mehrabian, Audrey Durand, Branislav Kveton

We propose $\tt RandUCB$, a bandit strategy that uses theoretically derived confidence intervals similar to upper confidence bound (UCB) algorithms, but akin to Thompson sampling (TS), uses randomization to trade off exploration and exploitation. In the $K$-armed bandit setting, we show that there are infinitely many variants of $\tt RandUCB$, all of which achieve the minimax-optimal $\widetilde{O}(\sqrt{K T})$ regret after $T$ rounds. Moreover, in a specific multi-armed bandit setting, we show that both UCB and TS can be recovered as special cases of $\tt RandUCB.$ For structured bandits, where each arm is associated with a $d$-dimensional feature vector and rewards are distributed according to a linear or generalized linear model, we prove that $\tt RandUCB$ achieves the minimax-optimal $\widetilde{O}(d \sqrt{T})$ regret even in the case of infinite arms. We demonstrate the practical effectiveness of $\tt RandUCB$ with experiments in both the multi-armed and structured bandit settings. Our results illustrate that $\tt RandUCB$ matches the empirical performance of TS while obtaining the theoretically optimal regret bounds of UCB algorithms, thus achieving the best of both worlds.

##### Temporal Regularization in Markov Decision Process

Nov 01, 2018
Pierre Thodoroff, Audrey Durand, Joelle Pineau, Doina Precup

Several applications of Reinforcement Learning suffer from instability due to high variance. This is especially prevalent in high dimensional domains. Regularization is a commonly used technique in machine learning to reduce variance, at the cost of introducing some bias. Most existing regularization techniques focus on spatial (perceptual) regularization. Yet in reinforcement learning, due to the nature of the Bellman equation, there is an opportunity to also exploit temporal regularization based on smoothness in value estimates over trajectories. This paper explores a class of methods for temporal regularization. We formally characterize the bias induced by this technique using Markov chain concepts. We illustrate the various characteristics of temporal regularization via a sequence of simple discrete and continuous MDPs, and show that the technique provides improvement even in high-dimensional Atari games.

* Published as a conference paper at NIPS 2018
##### MRPC: An R package for accurate inference of causal graphs

Jun 05, 2018
We present MRPC, an R package that learns causal graphs with improved accuracy over existing packages, such as pcalg and bnlearn. Our algorithm builds on the powerful PC algorithm, the canonical algorithm in computer science for learning directed acyclic graphs. The improvement in accuracy results from online control of the false discovery rate (FDR) that reduces false positive edges, a more accurate approach to identifying v-structures (i.e., $T_1 \rightarrow T_2 \leftarrow T_3$), and robust estimation of the correlation matrix among nodes. For genomic data that contain genotypes and gene expression for each sample, MRPC incorporates the principle of Mendelian randomization to orient the edges. Our package can be applied to continuous and discrete data.