April 2, 2020

3042 words 15 mins read

Paper Group ANR 263

Efficient Deep Reinforcement Learning through Policy Transfer. Addressing target shift in zero-shot learning using grouped adversarial learning. Towards Zero-shot Learning for Automatic Phonemic Transcription. Cognitive Argumentation and the Suppression Task. A Multi-criteria Approach for Fast and Outlier-aware Representative Selection from Manifol …

Efficient Deep Reinforcement Learning through Policy Transfer


Title	Efficient Deep Reinforcement Learning through Policy Transfer
Authors	Tianpei Yang, Jianye Hao, Zhaopeng Meng, Zongzhang Zhang, Weixun Wang, Yujing Hu, Yingfeng Cheng, Changjie Fan, Zhaodong Wang, Jiajie Peng
Abstract	Transfer Learning (TL) has shown great potential to accelerate Reinforcement Learning (RL) by leveraging prior knowledge from past learned policies of relevant tasks. Existing transfer approaches either explicitly computes the similarity between tasks or select appropriate source policies to provide guided explorations for the target task. However, how to directly optimize the target policy by alternatively utilizing knowledge from appropriate source policies without explicitly measuring the similarity is currently missing. In this paper, we propose a novel Policy Transfer Framework (PTF) to accelerate RL by taking advantage of this idea. Our framework learns when and which source policy is the best to reuse for the target policy and when to terminate it by modeling multi-policy transfer as the option learning problem. PTF can be easily combined with existing deep RL approaches. Experimental results show it significantly accelerates the learning process and surpasses state-of-the-art policy transfer methods in terms of learning efficiency and final performance in both discrete and continuous action spaces.
Tasks	Transfer Learning
Published	2020-02-19
URL	https://arxiv.org/abs/2002.08037v1
PDF	https://arxiv.org/pdf/2002.08037v1.pdf
PWC	https://paperswithcode.com/paper/efficient-deep-reinforcement-learning-through
Repo
Framework

Addressing target shift in zero-shot learning using grouped adversarial learning


Title	Addressing target shift in zero-shot learning using grouped adversarial learning
Authors	Saneem Ahmed Chemmengath, Samarth Bharadwaj, Soumava Paul, Suranjana Samanta, Karthik Sankaranarayanan
Abstract	In this paper, we present a new paradigm to zero-shot learning (ZSL) that is trained by utilizing additional information (such as attribute-class mapping) for specific set of unseen classes. We conjecture that such additional information about unseen classes is more readily available than unsupervised image sets. Further, on close examination of the underlying attribute predictors of popular ZSL algorithms, we find that they often leverage attribute correlations to make predictions. While attribute correlations that remain intact in the unseen classes (test) benefit the prediction of difficult attributes, change in correlations can have an adverse effect on ZSL performance. For example, detecting an attribute ‘brown’ may be the same as detecting ‘fur’ over an animals’ image dataset captured in the tropics. However, such a model might fail on unseen images of Arctic animals. To address this effect, termed target-shift in ZSL, we utilize our proposed framework to design grouped adversarial learning. We introduce grouping of attributes to enable the model to continue to benefit from useful correlations, while restricting cross-group correlations that may be harmful for generalization. Our analysis shows that it is possible to not only constrain the model from leveraging unwanted correlations, but also adjust them to specific test setting using only the additional information (the already available attribute-class mapping). We show empirical results for zero-shot predictions on standard benchmark datasets, namely, aPY, AwA2, SUN and CUB datasets. We further introduce to the research community, a new experimental train-test split that maximizes target-shift to further study its effects.
Tasks	Zero-Shot Learning
Published	2020-03-02
URL	https://arxiv.org/abs/2003.00845v1
PDF	https://arxiv.org/pdf/2003.00845v1.pdf
PWC	https://paperswithcode.com/paper/addressing-target-shift-in-zero-shot-learning
Repo
Framework

Towards Zero-shot Learning for Automatic Phonemic Transcription


Title	Towards Zero-shot Learning for Automatic Phonemic Transcription
Authors	Xinjian Li, Siddharth Dalmia, David R. Mortensen, Juncheng Li, Alan W Black, Florian Metze
Abstract	Automatic phonemic transcription tools are useful for low-resource language documentation. However, due to the lack of training sets, only a tiny fraction of languages have phonemic transcription tools. Fortunately, multilingual acoustic modeling provides a solution given limited audio training data. A more challenging problem is to build phonemic transcribers for languages with zero training data. The difficulty of this task is that phoneme inventories often differ between the training languages and the target language, making it infeasible to recognize unseen phonemes. In this work, we address this problem by adopting the idea of zero-shot learning. Our model is able to recognize unseen phonemes in the target language without any training data. In our model, we decompose phonemes into corresponding articulatory attributes such as vowel and consonant. Instead of predicting phonemes directly, we first predict distributions over articulatory attributes, and then compute phoneme distributions with a customized acoustic model. We evaluate our model by training it using 13 languages and testing it using 7 unseen languages. We find that it achieves 7.7% better phoneme error rate on average over a standard multilingual model.
Tasks	Zero-Shot Learning
Published	2020-02-26
URL	https://arxiv.org/abs/2002.11781v1
PDF	https://arxiv.org/pdf/2002.11781v1.pdf
PWC	https://paperswithcode.com/paper/towards-zero-shot-learning-for-automatic
Repo
Framework

Cognitive Argumentation and the Suppression Task


Title	Cognitive Argumentation and the Suppression Task
Authors	Emmanuelle-Anna Dietz Saldanha, Antonis Kakas
Abstract	This paper addresses the challenge of modeling human reasoning, within a new framework called Cognitive Argumentation. This framework rests on the assumption that human logical reasoning is inherently a process of dialectic argumentation and aims to develop a cognitive model for human reasoning that is computational and implementable. To give logical reasoning a human cognitive form the framework relies on cognitive principles, based on empirical and theoretical work in Cognitive Science, to suitably adapt a general and abstract framework of computational argumentation from AI. The approach of Cognitive Argumentation is evaluated with respect to Byrne’s suppression task, where the aim is not only to capture the suppression effect between different groups of people but also to account for the variation of reasoning within each group. Two main cognitive principles are particularly important to capture human conditional reasoning that explain the participants’ responses: (i) the interpretation of a condition within a conditional as sufficient and/or necessary and (ii) the mode of reasoning either as predictive or explanatory. We argue that Cognitive Argumentation provides a coherent and cognitively adequate model for human conditional reasoning that allows a natural distinction between definite and plausible conclusions, exhibiting the important characteristics of context-sensitive and defeasible reasoning.
Tasks
Published	2020-02-24
URL	https://arxiv.org/abs/2002.10149v1
PDF	https://arxiv.org/pdf/2002.10149v1.pdf
PWC	https://paperswithcode.com/paper/cognitive-argumentation-and-the-suppression
Repo
Framework

A Multi-criteria Approach for Fast and Outlier-aware Representative Selection from Manifolds


Title	A Multi-criteria Approach for Fast and Outlier-aware Representative Selection from Manifolds
Authors	Mahlagha Sedghi, George Atia, Michael Georgiopoulos
Abstract	The problem of representative selection amounts to sampling few informative exemplars from large datasets. This paper presents MOSAIC, a novel representative selection approach from high-dimensional data that may exhibit non-linear structures. Resting upon a novel quadratic formulation, Our method advances a multi-criteria selection approach that maximizes the global representation power of the sampled subset, ensures diversity, and rejects disruptive information by effectively detecting outliers. Through theoretical analyses we characterize the obtained sketch and reveal that the sampled representatives maximize a well-defined notion of data coverage in a transformed space. In addition, we present a highly scalable randomized implementation of the proposed algorithm shown to bring about substantial speedups. MOSAIC’s superiority in achieving the desired characteristics of a representative subset all at once while exhibiting remarkable robustness to various outlier types is demonstrated via extensive experiments conducted on both real and synthetic data with comparisons to state-of-the-art algorithms.
Tasks
Published	2020-03-12
URL	https://arxiv.org/abs/2003.05989v1
PDF	https://arxiv.org/pdf/2003.05989v1.pdf
PWC	https://paperswithcode.com/paper/a-multi-criteria-approach-for-fast-and
Repo
Framework

Variational Item Response Theory: Fast, Accurate, and Expressive


Title	Variational Item Response Theory: Fast, Accurate, and Expressive
Authors	Mike Wu, Richard L. Davis, Benjamin W. Domingue, Chris Piech, Noah Goodman
Abstract	Item Response Theory (IRT) is a ubiquitous model for understanding humans based on their responses to questions, used in fields as diverse as education, medicine and psychology. Large modern datasets offer opportunities to capture more nuances in human behavior, potentially improving test scoring and better informing public policy. Yet larger datasets pose a difficult speed / accuracy challenge to contemporary algorithms for fitting IRT models. We introduce a variational Bayesian inference algorithm for IRT, and show that it is fast and scaleable without sacrificing accuracy. Using this inference approach we then extend classic IRT with expressive Bayesian models of responses. Applying this method to five large-scale item response datasets from cognitive science and education yields higher log likelihoods and improvements in imputing missing data. The algorithm implementation is open-source, and easily usable.
Tasks	Bayesian Inference
Published	2020-02-01
URL	https://arxiv.org/abs/2002.00276v2
PDF	https://arxiv.org/pdf/2002.00276v2.pdf
PWC	https://paperswithcode.com/paper/variational-item-response-theory-fast
Repo
Framework

Balancing Cost and Benefit with Tied-Multi Transformers


Title	Balancing Cost and Benefit with Tied-Multi Transformers
Authors	Raj Dabre, Raphael Rubino, Atsushi Fujita
Abstract	We propose and evaluate a novel procedure for training multiple Transformers with tied parameters which compresses multiple models into one enabling the dynamic choice of the number of encoder and decoder layers during decoding. In sequence-to-sequence modeling, typically, the output of the last layer of the N-layer encoder is fed to the M-layer decoder, and the output of the last decoder layer is used to compute loss. Instead, our method computes a single loss consisting of NxM losses, where each loss is computed from the output of one of the M decoder layers connected to one of the N encoder layers. Such a model subsumes NxM models with different number of encoder and decoder layers, and can be used for decoding with fewer than the maximum number of encoder and decoder layers. We then propose a mechanism to choose a priori the number of encoder and decoder layers for faster decoding, and also explore recurrent stacking of layers and knowledge distillation for model compression. We present a cost-benefit analysis of applying the proposed approaches for neural machine translation and show that they reduce decoding costs while preserving translation quality.
Tasks	Machine Translation, Model Compression
Published	2020-02-20
URL	https://arxiv.org/abs/2002.08614v1
PDF	https://arxiv.org/pdf/2002.08614v1.pdf
PWC	https://paperswithcode.com/paper/balancing-cost-and-benefit-with-tied-multi-1
Repo
Framework

Lightweight Convolutional Representations for On-Device Natural Language Processing


Title	Lightweight Convolutional Representations for On-Device Natural Language Processing
Authors	Shrey Desai, Geoffrey Goh, Arun Babu, Ahmed Aly
Abstract	The increasing computational and memory complexities of deep neural networks have made it difficult to deploy them on low-resource electronic devices (e.g., mobile phones, tablets, wearables). Practitioners have developed numerous model compression methods to address these concerns, but few have condensed input representations themselves. In this work, we propose a fast, accurate, and lightweight convolutional representation that can be swapped into any neural model and compressed significantly (up to 32x) with a negligible reduction in performance. In addition, we show gains over recurrent representations when considering resource-centric metrics (e.g., model file size, latency, memory usage) on a Samsung Galaxy S9.
Tasks	Model Compression
Published	2020-02-04
URL	https://arxiv.org/abs/2002.01535v1
PDF	https://arxiv.org/pdf/2002.01535v1.pdf
PWC	https://paperswithcode.com/paper/lightweight-convolutional-representations-for
Repo
Framework

An Overview and Case Study of the Clinical AI Model Development Life Cycle for Healthcare Systems


Title	An Overview and Case Study of the Clinical AI Model Development Life Cycle for Healthcare Systems
Authors	Charles Lu, Julia Strout, Romane Gauriau, Brad Wright, Fabiola Bezerra De Carvalho Marcruz, Varun Buch, Katherine Andriole
Abstract	Healthcare is one of the most promising areas for machine learning models to make a positive impact. However, successful adoption of AI-based systems in healthcare depends on engaging and educating stakeholders from diverse backgrounds about the development process of AI models. We present a broadly accessible overview of the development life cycle of clinical AI models that is general enough to be adapted to most machine learning projects, and then give an in-depth case study of the development process of a deep learning based system to detect aortic aneurysms in Computed Tomography (CT) exams. We hope other healthcare institutions and clinical practitioners find the insights we share about the development process useful in informing their own model development efforts and to increase the likelihood of successful deployment and integration of AI in healthcare.
Tasks	Computed Tomography (CT)
Published	2020-03-02
URL	https://arxiv.org/abs/2003.07678v3
PDF	https://arxiv.org/pdf/2003.07678v3.pdf
PWC	https://paperswithcode.com/paper/an-overview-and-case-study-of-the-clinical-ai
Repo
Framework

MaskFlownet: Asymmetric Feature Matching with Learnable Occlusion Mask


Title	MaskFlownet: Asymmetric Feature Matching with Learnable Occlusion Mask
Authors	Shengyu Zhao, Yilun Sheng, Yue Dong, Eric I-Chao Chang, Yan Xu
Abstract	Feature warping is a core technique in optical flow estimation; however, the ambiguity caused by occluded areas during warping is a major problem that remains unsolved. In this paper, we propose an asymmetric occlusion-aware feature matching module, which can learn a rough occlusion mask that filters useless (occluded) areas immediately after feature warping without any explicit supervision. The proposed module can be easily integrated into end-to-end network architectures and enjoys performance gains while introducing negligible computational cost. The learned occlusion mask can be further fed into a subsequent network cascade with dual feature pyramids with which we achieve state-of-the-art performance. At the time of submission, our method, called MaskFlownet, surpasses all published optical flow methods on the MPI Sintel, KITTI 2012 and 2015 benchmarks. Code is available at https://github.com/microsoft/MaskFlownet.
Tasks	Optical Flow Estimation
Published	2020-03-24
URL	https://arxiv.org/abs/2003.10955v1
PDF	https://arxiv.org/pdf/2003.10955v1.pdf
PWC	https://paperswithcode.com/paper/maskflownet-asymmetric-feature-matching-with
Repo
Framework

CNN Hyperparameter tuning applied to Iris Liveness Detection


Title	CNN Hyperparameter tuning applied to Iris Liveness Detection
Authors	Gabriela Y. Kimura, Diego R. Lucio, Alceu S. Britto Jr., David Menotti
Abstract	The iris pattern has significantly improved the biometric recognition field due to its high level of stability and uniqueness. Such physical feature has played an important role in security and other related areas. However, presentation attacks, also known as spoofing techniques, can be used to bypass the biometric system with artifacts such as printed images, artificial eyes, and textured contact lenses. To improve the security of these systems, many liveness detection methods have been proposed, and the first Internacional Iris Liveness Detection competition was launched in 2013 to evaluate their effectiveness. In this paper, we propose a hyperparameter tuning of the CASIA algorithm, submitted by the Chinese Academy of Sciences to the third competition of Iris Liveness Detection, in 2017. The modifications proposed promoted an overall improvement, with an 8.48% Attack Presentation Classification Error Rate (APCER) and 0.18% Bonafide Presentation Classification Error Rate (BPCER) for the evaluation of the combined datasets. Other threshold values were evaluated in an attempt to reduce the trade-off between the APCER and the BPCER on the evaluated datasets and worked out successfully.
Tasks
Published	2020-02-12
URL	https://arxiv.org/abs/2003.00833v1
PDF	https://arxiv.org/pdf/2003.00833v1.pdf
PWC	https://paperswithcode.com/paper/cnn-hyperparameter-tuning-applied-to-iris
Repo
Framework

Performative Prediction


Title	Performative Prediction
Authors	Juan C. Perdomo, Tijana Zrnic, Celestine Mendler-Dünner, Moritz Hardt
Abstract	When predictions support decisions they may influence the outcome they aim to predict. We call such predictions performative; the prediction influences the target. Performativity is a well-studied phenomenon in policy-making that has so far been neglected in supervised learning. When ignored, performativity surfaces as undesirable distribution shift, routinely addressed with retraining. We develop a risk minimization framework for performative prediction bringing together concepts from statistics, game theory, and causality. A conceptual novelty is an equilibrium notion we call performative stability. Performative stability implies that the predictions are calibrated not against past outcomes, but against the future outcomes that manifest from acting on the prediction. Our main results are necessary and sufficient conditions for the convergence of retraining to a performatively stable point of nearly minimal loss. In full generality, performative prediction strictly subsumes the setting known as strategic classification. We thus also give the first sufficient conditions for retraining to overcome strategic feedback effects.
Tasks
Published	2020-02-16
URL	https://arxiv.org/abs/2002.06673v1
PDF	https://arxiv.org/pdf/2002.06673v1.pdf
PWC	https://paperswithcode.com/paper/performative-prediction
Repo
Framework

The Archimedean trap: Why traditional reinforcement learning will probably not yield AGI


Title	The Archimedean trap: Why traditional reinforcement learning will probably not yield AGI
Authors	Samuel Allen Alexander
Abstract	After generalizing the Archimedean property of real numbers in such a way as to make it adaptable to non-numeric structures, we demonstrate that the real numbers cannot be used to accurately measure non-Archimedean structures. We argue that, since an agent with Artificial General Intelligence (AGI) should have no problem engaging in tasks that inherently involve non-Archimedean rewards, and since traditional reinforcement learning rewards are real numbers, therefore traditional reinforcement learning cannot lead to AGI. We indicate two possible ways traditional reinforcement learning could be altered to remove this roadblock.
Tasks
Published	2020-02-15
URL	https://arxiv.org/abs/2002.10221v1
PDF	https://arxiv.org/pdf/2002.10221v1.pdf
PWC	https://paperswithcode.com/paper/the-archimedean-trap-why-traditional
Repo
Framework

Evading Deepfake-Image Detectors with White- and Black-Box Attacks


Title	Evading Deepfake-Image Detectors with White- and Black-Box Attacks
Authors	Nicholas Carlini, Hany Farid
Abstract	It is now possible to synthesize highly realistic images of people who don’t exist. Such content has, for example, been implicated in the creation of fraudulent social-media profiles responsible for dis-information campaigns. Significant efforts are, therefore, being deployed to detect synthetically-generated content. One popular forensic approach trains a neural network to distinguish real from synthetic content. We show that such forensic classifiers are vulnerable to a range of attacks that reduce the classifier to near-0% accuracy. We develop five attack case studies on a state-of-the-art classifier that achieves an area under the ROC curve (AUC) of 0.95 on almost all existing image generators, when only trained on one generator. With full access to the classifier, we can flip the lowest bit of each pixel in an image to reduce the classifier’s AUC to 0.0005; perturb 1% of the image area to reduce the classifier’s AUC to 0.08; or add a single noise pattern in the synthesizer’s latent space to reduce the classifier’s AUC to 0.17. We also develop a black-box attack that, with no access to the target classifier, reduces the AUC to 0.22. These attacks reveal significant vulnerabilities of certain image-forensic classifiers.
Tasks	Face Swapping
Published	2020-04-01
URL	https://arxiv.org/abs/2004.00622v1
PDF	https://arxiv.org/pdf/2004.00622v1.pdf
PWC	https://paperswithcode.com/paper/evading-deepfake-image-detectors-with-white
Repo
Framework

Articulation-aware Canonical Surface Mapping


Title	Articulation-aware Canonical Surface Mapping
Authors	Nilesh Kulkarni, Abhinav Gupta, David F. Fouhey, Shubham Tulsiani
Abstract	We tackle the tasks of: 1) predicting a Canonical Surface Mapping (CSM) that indicates the mapping from 2D pixels to corresponding points on a canonical template shape, and 2) inferring the articulation and pose of the template corresponding to the input image. While previous approaches rely on keypoint supervision for learning, we present an approach that can learn without such annotations. Our key insight is that these tasks are geometrically related, and we can obtain supervisory signal via enforcing consistency among the predictions. We present results across a diverse set of animal object categories, showing that our method can learn articulation and CSM prediction from image collections using only foreground mask labels for training. We empirically show that allowing articulation helps learn more accurate CSM prediction, and that enforcing the consistency with predicted CSM is similarly critical for learning meaningful articulation.
Tasks
Published	2020-04-01
URL	https://arxiv.org/abs/2004.00614v1
PDF	https://arxiv.org/pdf/2004.00614v1.pdf
PWC	https://paperswithcode.com/paper/articulation-aware-canonical-surface-mapping
Repo
Framework