Tingting Zhao

Tingting Zhao, Ying Wang, Wei Sun et al.

Deep reinforcement learning (DRL) breaks through the bottlenecks of traditional reinforcement learning (RL) with the help of the perception capability of deep learning and has been widely applied in real-world problems.While model-free RL, as a class of efficient DRL methods, performs the learning of state representations simultaneously with policy learning in an end-to-end manner when facing large-scale continuous state and action spaces. However, training such a large policy model requires a large number of trajectory samples and training time. On the other hand, the learned policy often fails to generalize to large-scale action spaces, especially for the continuous action spaces. To address this issue, in this paper we propose an efficient policy learning method in latent state and action spaces. More specifically, we extend the idea of state representations to action representations for better policy generalization capability. Meanwhile, we divide the whole learning task into learning with the large-scale representation models in an unsupervised manner and learning with the small-scale policy model in the RL manner.The small policy model facilitates policy learning, while not sacrificing generalization and expressiveness via the large representation model. Finally,the effectiveness of the proposed method is demonstrated by MountainCar,CarRacing and Cheetah experiments.

Thomas Cook, Harsh Vardhan Dubey, Ji Ah Lee et al.

We consider the problem of sequential multiple hypothesis testing with nontrivial data collection costs. This problem appears, for example, when conducting biological experiments to identify differentially expressed genes of a disease process. This work builds on the generalized $α$-investing framework which enables control of the false discovery rate in a sequential testing setting. We make a theoretical analysis of the long term asymptotic behavior of $α$-wealth which motivates a consideration of sample size in the $α$-investing decision rule. Posing the testing process as a game with nature, we construct a decision rule that optimizes the expected $α$-wealth reward (ERO) and provides an optimal sample size for each test. Empirical results show that a cost-aware ERO decision rule correctly rejects more false null hypotheses than other methods for $n=1$ where $n$ is the sample size. When the sample size is not fixed cost-aware ERO uses a prior on the null hypothesis to adaptively allocate of the sample budget to each test. We extend cost-aware ERO investing to finite-horizon testing which enables the decision rule to allocate samples in a non-myopic manner. Finally, empirical tests on real data sets from biological experiments show that cost-aware ERO balances the allocation of samples to an individual test against the allocation of samples across multiple tests.

Yuan Wang, Huiling Song, Peng Huo et al.

Extreme multi-label text classification (XMTC) refers to the problem of tagging a given text with the most relevant subset of labels from a large label set. A majority of labels only have a few training instances due to large label dimensionality in XMTC. To solve this data sparsity issue, most existing XMTC methods take advantage of fixed label clusters obtained in early stage to balance performance on tail labels and head labels. However, such label clusters provide static and coarse-grained semantic scope for every text, which ignores distinct characteristics of different texts and has difficulties modelling accurate semantics scope for texts with tail labels. In this paper, we propose a novel framework TReaderXML for XMTC, which adopts dynamic and fine-grained semantic scope from teacher knowledge for individual text to optimize text conditional prior category semantic ranges. TReaderXML dynamically obtains teacher knowledge for each text by similar texts and hierarchical label information in training sets to release the ability of distinctly fine-grained label-oriented semantic scope. Then, TReaderXML benefits from a novel dual cooperative network that firstly learns features of a text and its corresponding label-oriented semantic scope by parallel Encoding Module and Reading Module, secondly embeds two parts by Interaction Module to regularize the text's representation by dynamic and fine-grained label-oriented semantic scope, and finally find target labels by Prediction Module. Experimental results on three XMTC benchmark datasets show that our method achieves new state-of-the-art results and especially performs well for severely imbalanced and sparse datasets.

Tingting Zhao, Shubo Tian, Jordan Daly et al.

For an approaching disaster, the tracking of time-sensitive critical information such as hurricane evacuation notices is challenging in the United States. These notices are issued and distributed rapidly by numerous local authorities that may spread across multiple states. They often undergo frequent updates and are distributed through diverse online portals lacking standard formats. In this study, we developed an approach to timely detect and track the locally issued hurricane evacuation notices. The text data were collected mainly with a spatially targeted web scraping method. They were manually labeled and then classified using natural language processing techniques with deep learning models. The classification of mandatory evacuation notices achieved a high accuracy (recall = 96%). We used Hurricane Ian (2022) to illustrate how real-time evacuation notices extracted from local government sources could be redistributed with a Web GIS system. Our method applied to future hurricanes provides live data for situation awareness to higher-level government agencies and news media. The archived data helps scholars to study government responses toward weather warnings and individual behaviors influenced by evacuation history. The framework may be applied to other types of disasters for rapid and targeted retrieval, classification, redistribution, and archiving of real-time government orders and notifications.

Tingting Zhao, Zifeng Wang, Aria Masoomi et al.

Lifelong Learning (LL) refers to the ability to continually learn and solve new problems with incremental available information over time while retaining previous knowledge. Much attention has been given lately to Supervised Lifelong Learning (SLL) with a stream of labelled data. In contrast, we focus on resolving challenges in Unsupervised Lifelong Learning (ULL) with streaming unlabelled data when the data distribution and the unknown class labels evolve over time. Bayesian framework is natural to incorporate past knowledge and sequentially update the belief with new data. We develop a fully Bayesian inference framework for ULL with a novel end-to-end Deep Bayesian Unsupervised Lifelong Learning (DBULL) algorithm, which can progressively discover new clusters without forgetting the past with unlabelled data while learning latent representations. To efficiently maintain past knowledge, we develop a novel knowledge preservation mechanism via sufficient statistics of the latent representation for raw data. To detect the potential new clusters on the fly, we develop an automatic cluster discovery and redundancy removal strategy in our inference inspired by Nonparametric Bayesian statistics techniques. We demonstrate the effectiveness of our approach using image and text corpora benchmark datasets in both LL and batch settings.

Zhaowei Zhu, Xiang Lan, Tingting Zhao et al.

Cardiovascular disease is a major threat to health and one of the primary causes of death globally. The 12-lead ECG is a cheap and commonly accessible tool to identify cardiac abnormalities. Early and accurate diagnosis will allow early treatment and intervention to prevent severe complications of cardiovascular disease. In the PhysioNet/Computing in Cardiology Challenge 2020, our objective is to develop an algorithm that automatically identifies 27 ECG abnormalities from 12-lead ECG recordings.

Tingting Zhao, Hang Zhang, Jacob Spoelstra

We worked with Nestle SHIELD (Skin Health, Innovation, Education, and Longevity Development, NSH) to develop a deep learning model that is able to assess acne severity from selfie images as accurate as dermatologists. The model was deployed as a mobile application, providing patients an easy way to assess and track the progress of their acne treatment. NSH acquired 4,700 selfie images for this study and recruited 11 internal dermatologists to label them in five categories: 1-Clear, 2- Almost Clear, 3-Mild, 4-Moderate, 5-Severe. Using OpenCV to detect facial landmarks we cut specific skin patches from the selfie images in order to minimize irrelevant background. We then applied a transfer learning approach by extracting features from the patches using a ResNet 152 pre-trained model, followed by a fully connected layer trained to approximate the desired severity rating. To address the problem of spatial sensitivity of CNN models, we introduce a new image rolling data augmentation approach, effectively causing acne lesions appeared in more locations in the training images. Our results demonstrate that this approach improved the generalization of the CNN model, outperforming more than half of the panel of human dermatologists on test images. To our knowledge, this is the first deep learning-based solution for acne assessment using selfie images.

Tingting Zhao, Zifeng Wang, Aria Masoomi et al.

We develop a data driven approach to perform clustering and end-to-end feature learning simultaneously for streaming data that can adaptively detect novel clusters in emerging data. Our approach, Adaptive Nonparametric Variational Autoencoder (AdapVAE), learns the cluster membership through a Bayesian Nonparametric (BNP) modeling framework with Deep Neural Networks (DNNs) for feature learning. We develop a joint online variational inference algorithm to learn feature representations and clustering assignments simultaneously via iteratively optimizing the Evidence Lower Bound (ELBO). We resolve the catastrophic forgetting \citep{kirkpatrick2017overcoming} challenges with streaming data by adopting generative samples from the trained AdapVAE using previous data, which avoids the need of storing and reusing past data. We demonstrate the advantages of our model including adaptive novel cluster detection without discarding useful information learned from past data, high quality sample generation and comparable clustering performance as end-to-end batch mode clustering methods on both image and text corpora benchmark datasets.

Tingting Zhao, Alexandre Bouchard-Côté

Sampling the parameters of high-dimensional Continuous Time Markov Chains (CTMC) is a challenging problem with important applications in many fields of applied statistics. In this work a recently proposed type of non-reversible rejection-free Markov Chain Monte Carlo (MCMC) sampler, the Bouncy Particle Sampler (BPS), is brought to bear to this problem. BPS has demonstrated its favorable computational efficiency compared with state-of-the-art MCMC algorithms, however to date applications to real-data scenario were scarce. An important aspect of the practical implementation of BPS is the simulation of event times. Default implementations use conservative thinning bounds. Such bounds can slow down the algorithm and limit the computational performance. Our paper develops an algorithm with an exact analytical solution to the random event times in the context of CTMCs. Our local version of BPS algorithm takes advantage of the sparse structure in the target factor graph and we also provide a framework for assessing the computational complexity of local BPS algorithms.

Guangyu Zhu, Tingting Zhao

Feature selection is central to contemporary high-dimensional data analysis. Grouping structure among features arises naturally in various scientific problems. Many methods have been proposed to incorporate the grouping structure information into feature selection. However, these methods are normally restricted to a linear regression setting. To relax the linear constraint, we combine the deep neural networks (DNNs) with the recent Knockoffs technique, which has been successful in an individual feature selection context. We propose Deep-gKnock (Deep group-feature selection using Knockoffs) as a methodology for model interpretation and dimension reduction. Deep-gKnock performs model-free group-feature selection by controlling group-wise False Discovery Rate (gFDR). Our method improves the interpretability and reproducibility of DNNs. Experimental results on both synthetic and real data demonstrate that our method achieves superior power and accurate gFDR control compared with state-of-the-art methods.

Albert Steppi, Jinchan Qu, Minjing Tao et al.

The increasing volume of scientific publications and grant proposals has generated an unprecedentedly high workload to scientific communities. Consequently, review quality has been decreasing and review outcomes have become less correlated with the real merits of the papers and proposals. A novel distributed peer review (DPR) approach has recently been proposed to address these issues. The new approach assigns principal investigators (PIs) who submitted proposals (or papers) to the same program as reviewers. Each PI reviews and ranks a small number (such as seven) of other PIs' proposals. The individual rankings are then used to estimate a global ranking of all proposals using the Modified Borda Count (MBC). In this study, we perform simulation studies to investigate several parameters important for the decision making when adopting this new approach. We also propose a new method called Concordance Index-based Global Ranking (CIGR) to estimate global ranking from individual rankings. An efficient simulated annealing algorithm is designed to search the optimal Concordance Index (CI). Moreover, we design a new balanced review assignment procedure, which can result in significantly better performance for both MBC and CIGR methods. We found that CIGR performs better than MBC when the review quality is relatively high. As review quality and review difficulty are tightly correlated, we constructed a boundary in the space of review quality vs review difficulty that separates the CIGR-superior and MBC-superior regions. Finally, we propose a multi-stage DPR strategy based on CIGR, which has the potential to substantially improve the overall review performance while reducing the review workload.

Norikazu Sugimoto, Voot Tangkaratt, Thijs Wensveen et al.

In this study, we show that a movement policy can be improved efficiently using the previous experiences of a real robot. Reinforcement Learning (RL) is becoming a popular approach to acquire a nonlinear optimal policy through trial and error. However, it is considered very difficult to apply RL to real robot control since it usually requires many learning trials. Such trials cannot be executed in real environments because unrealistic time is necessary and the real system's durability is limited. Therefore, in this study, instead of executing many learning trials, we propose to use a recently developed RL algorithm, importance-weighted PGPE, by which the robot can efficiently reuse previously sampled data to improve it's policy parameters. We apply importance-weighted PGPE to CB-i, our real humanoid robot, and show that it can learn a target reaching movement and a cart-pole swing up movement in a real environment without using any prior knowledge of the task or any carefully designed initial trajectory.

Syogo Mori, Voot Tangkaratt, Tingting Zhao et al.

The goal of reinforcement learning (RL) is to let an agent learn an optimal control policy in an unknown environment so that future expected rewards are maximized. The model-free RL approach directly learns the policy based on data samples. Although using many samples tends to improve the accuracy of policy learning, collecting a large number of samples is often expensive in practice. On the other hand, the model-based RL approach first estimates the transition model of the environment and then learns the policy based on the estimated transition model. Thus, if the transition model is accurately learned from a small amount of data, the model-based approach can perform better than the model-free approach. In this paper, we propose a novel model-based RL method by combining a recently proposed model-free policy search method called policy gradients with parameter-based exploration and the state-of-the-art transition model estimator called least-squares conditional density estimation. Through experiments, we demonstrate the practical usefulness of the proposed method.

Tingting Zhao, Hirotaka Hachiya, Voot Tangkaratt et al.

The policy gradient approach is a flexible and powerful reinforcement learning method particularly for problems with continuous actions such as robot control. A common challenge in this scenario is how to reduce the variance of policy gradient estimates for reliable policy updates. In this paper, we combine the following three ideas and give a highly effective policy gradient method: (a) the policy gradients with parameter based exploration, which is a recently proposed policy search method with low variance of gradient estimates, (b) an importance sampling technique, which allows us to reuse previously gathered data in a consistent way, and (c) an optimal baseline, which minimizes the variance of gradient estimates with their unbiasedness being maintained. For the proposed method, we give theoretical analysis of the variance of gradient estimates and show its usefulness through extensive experiments.