Paper Group ANR 269
Multi-Source Neural Machine Translation with Data Augmentation. Less is More: Surgical Phase Recognition with Less Annotations through Self-Supervised Pre-training of CNN-LSTM Networks. On the Transferability of Representations in Neural Networks Between Datasets and Tasks. Towards Understanding the Generalization Bias of Two Layer Convolutional Li …
Multi-Source Neural Machine Translation with Data Augmentation
| Title | Multi-Source Neural Machine Translation with Data Augmentation |
| Authors | Yuta Nishimura, Katsuhito Sudoh, Graham Neubig, Satoshi Nakamura |
| Abstract | Multi-source translation systems translate from multiple languages to a single target language. By using information from these multiple sources, these systems achieve large gains in accuracy. To train these systems, it is necessary to have corpora with parallel text in multiple sources and the target language. However, these corpora are rarely complete in practice due to the difficulty of providing human translations in all of the relevant languages. In this paper, we propose a data augmentation approach to fill such incomplete parts using multi-source neural machine translation (NMT). In our experiments, results varied over different language combinations but significant gains were observed when using a source language similar to the target language. |
| Tasks | Data Augmentation, Machine Translation |
| Published | 2018-10-16 |
| URL | http://arxiv.org/abs/1810.06826v2 |
| http://arxiv.org/pdf/1810.06826v2.pdf | |
| PWC | https://paperswithcode.com/paper/multi-source-neural-machine-translation-with-1 |
| Repo | |
| Framework | |
Less is More: Surgical Phase Recognition with Less Annotations through Self-Supervised Pre-training of CNN-LSTM Networks
| Title | Less is More: Surgical Phase Recognition with Less Annotations through Self-Supervised Pre-training of CNN-LSTM Networks |
| Authors | Gaurav Yengera, Didier Mutter, Jacques Marescaux, Nicolas Padoy |
| Abstract | Real-time algorithms for automatically recognizing surgical phases are needed to develop systems that can provide assistance to surgeons, enable better management of operating room (OR) resources and consequently improve safety within the OR. State-of-the-art surgical phase recognition algorithms using laparoscopic videos are based on fully supervised training. This limits their potential for widespread application, since creation of manual annotations is an expensive process considering the numerous types of existing surgeries and the vast amount of laparoscopic videos available. In this work, we propose a new self-supervised pre-training approach based on the prediction of remaining surgery duration (RSD) from laparoscopic videos. The RSD prediction task is used to pre-train a convolutional neural network (CNN) and long short-term memory (LSTM) network in an end-to-end manner. Our proposed approach utilizes all available data and reduces the reliance on annotated data, thereby facilitating the scaling up of surgical phase recognition algorithms to different kinds of surgeries. Additionally, we present EndoN2N, an end-to-end trained CNN-LSTM model for surgical phase recognition and evaluate the performance of our approach on a dataset of 120 Cholecystectomy laparoscopic videos (Cholec120). This work also presents the first systematic study of self-supervised pre-training approaches to understand the amount of annotations required for surgical phase recognition. Interestingly, the proposed RSD pre-training approach leads to performance improvement even when all the training data is manually annotated and outperforms the single pre-training approach for surgical phase recognition presently published in the literature. It is also observed that end-to-end training of CNN-LSTM networks boosts surgical phase recognition performance. |
| Tasks | |
| Published | 2018-05-22 |
| URL | http://arxiv.org/abs/1805.08569v1 |
| http://arxiv.org/pdf/1805.08569v1.pdf | |
| PWC | https://paperswithcode.com/paper/less-is-more-surgical-phase-recognition-with |
| Repo | |
| Framework | |
On the Transferability of Representations in Neural Networks Between Datasets and Tasks
| Title | On the Transferability of Representations in Neural Networks Between Datasets and Tasks |
| Authors | Haytham M. Fayek, Lawrence Cavedon, Hong Ren Wu |
| Abstract | Deep networks, composed of multiple layers of hierarchical distributed representations, tend to learn low-level features in initial layers and transition to high-level features towards final layers. Paradigms such as transfer learning, multi-task learning, and continual learning leverage this notion of generic hierarchical distributed representations to share knowledge across datasets and tasks. Herein, we study the layer-wise transferability of representations in deep networks across a few datasets and tasks and note some interesting empirical observations. |
| Tasks | Continual Learning, Multi-Task Learning, Transfer Learning |
| Published | 2018-11-29 |
| URL | http://arxiv.org/abs/1811.12273v1 |
| http://arxiv.org/pdf/1811.12273v1.pdf | |
| PWC | https://paperswithcode.com/paper/on-the-transferability-of-representations-in |
| Repo | |
| Framework | |
Towards Understanding the Generalization Bias of Two Layer Convolutional Linear Classifiers with Gradient Descent
| Title | Towards Understanding the Generalization Bias of Two Layer Convolutional Linear Classifiers with Gradient Descent |
| Authors | Yifan Wu, Barnabas Poczos, Aarti Singh |
| Abstract | A major challenge in understanding the generalization of deep learning is to explain why (stochastic) gradient descent can exploit the network architecture to find solutions that have good generalization performance when using high capacity models. We find simple but realistic examples showing that this phenomenon exists even when learning linear classifiers — between two linear networks with the same capacity, the one with a convolutional layer can generalize better than the other when the data distribution has some underlying spatial structure. We argue that this difference results from a combination of the convolution architecture, data distribution and gradient descent, all of which are necessary to be included in a meaningful analysis. We provide a general analysis of the generalization performance as a function of data distribution and convolutional filter size, given gradient descent as the optimization algorithm, then interpret the results using concrete examples. Experimental results show that our analysis is able to explain what happens in our introduced examples. |
| Tasks | |
| Published | 2018-02-13 |
| URL | http://arxiv.org/abs/1802.04420v2 |
| http://arxiv.org/pdf/1802.04420v2.pdf | |
| PWC | https://paperswithcode.com/paper/towards-understanding-the-generalization-bias |
| Repo | |
| Framework | |
HELOC Applicant Risk Performance Evaluation by Topological Hierarchical Decomposition
| Title | HELOC Applicant Risk Performance Evaluation by Topological Hierarchical Decomposition |
| Authors | Kyle Brown, Derek Doran, Ryan Kramer, Brad Reynolds |
| Abstract | Strong regulations in the financial industry mean that any decisions based on machine learning need to be explained. This precludes the use of powerful supervised techniques such as neural networks. In this study we propose a new unsupervised and semi-supervised technique known as the topological hierarchical decomposition (THD). This process breaks a dataset down into ever smaller groups, where groups are associated with a simplicial complex that approximate the underlying topology of a dataset. We apply THD to the FICO machine learning challenge dataset, consisting of anonymized home equity loan applications using the MAPPER algorithm to build simplicial complexes. We identify different groups of individuals unable to pay back loans, and illustrate how the distribution of feature values in a simplicial complex can be used to explain the decision to grant or deny a loan by extracting illustrative explanations from two THDs on the dataset. |
| Tasks | |
| Published | 2018-11-26 |
| URL | http://arxiv.org/abs/1811.10658v1 |
| http://arxiv.org/pdf/1811.10658v1.pdf | |
| PWC | https://paperswithcode.com/paper/heloc-applicant-risk-performance-evaluation |
| Repo | |
| Framework | |
An Extensive Review on Spectral Imaging in Biometric Systems: Challenges and Advancements
| Title | An Extensive Review on Spectral Imaging in Biometric Systems: Challenges and Advancements |
| Authors | Rumaisah Munir, Rizwan Ahmed Khan |
| Abstract | Spectral imaging has recently gained traction for face recognition in biometric systems. We investigate the merits of spectral imaging for face recognition and the current challenges that hamper the widespread deployment of spectral sensors for face recognition. The reliability of conventional face recognition systems operating in the visible range is compromised by illumination changes, pose variations and spoof attacks. Recent works have reaped the benefits of spectral imaging to counter these limitations in surveillance activities (defence, airport security checks, etc.). However, the implementation of this technology for biometrics, is still in its infancy due to multiple reasons. We present an overview of the existing work in the domain of spectral imaging for face recognition, different types of modalities and their assessment, availability of public databases for sake of reproducible research as well as evaluation of algorithms, and recent advancements in the field, such as, the use of deep learning-based methods for recognizing faces from spectral images. |
| Tasks | Face Recognition |
| Published | 2018-07-16 |
| URL | https://arxiv.org/abs/1807.05771v2 |
| https://arxiv.org/pdf/1807.05771v2.pdf | |
| PWC | https://paperswithcode.com/paper/an-extensive-review-on-spectral-imaging-in |
| Repo | |
| Framework | |
Acoustic and Textual Data Augmentation for Improved ASR of Code-Switching Speech
| Title | Acoustic and Textual Data Augmentation for Improved ASR of Code-Switching Speech |
| Authors | Emre Yılmaz, Henk van den Heuvel, David A. van Leeuwen |
| Abstract | In this paper, we describe several techniques for improving the acoustic and language model of an automatic speech recognition (ASR) system operating on code-switching (CS) speech. We focus on the recognition of Frisian-Dutch radio broadcasts where one of the mixed languages, namely Frisian, is an under-resourced language. In previous work, we have proposed several automatic transcription strategies for CS speech to increase the amount of available training speech data. In this work, we explore how the acoustic modeling (AM) can benefit from monolingual speech data belonging to the high-resourced mixed language. For this purpose, we train state-of-the-art AMs, which were ineffective due to lack of training data, on a significantly increased amount of CS speech and monolingual Dutch speech. Moreover, we improve the language model (LM) by creating code-switching text, which is in practice almost non-existent, by (1) generating text using recurrent LMs trained on the transcriptions of the training CS speech data, (2) adding the transcriptions of the automatically transcribed CS speech data and (3) translating Dutch text extracted from the transcriptions of a large Dutch speech corpora. We report significantly improved CS ASR performance due to the increase in the acoustic and textual training data. |
| Tasks | Data Augmentation, Language Modelling, Speech Recognition |
| Published | 2018-07-28 |
| URL | http://arxiv.org/abs/1807.10945v1 |
| http://arxiv.org/pdf/1807.10945v1.pdf | |
| PWC | https://paperswithcode.com/paper/acoustic-and-textual-data-augmentation-for |
| Repo | |
| Framework | |
Truncated Horizon Policy Search: Combining Reinforcement Learning & Imitation Learning
| Title | Truncated Horizon Policy Search: Combining Reinforcement Learning & Imitation Learning |
| Authors | Wen Sun, J. Andrew Bagnell, Byron Boots |
| Abstract | In this paper, we propose to combine imitation and reinforcement learning via the idea of reward shaping using an oracle. We study the effectiveness of the near-optimal cost-to-go oracle on the planning horizon and demonstrate that the cost-to-go oracle shortens the learner’s planning horizon as function of its accuracy: a globally optimal oracle can shorten the planning horizon to one, leading to a one-step greedy Markov Decision Process which is much easier to optimize, while an oracle that is far away from the optimality requires planning over a longer horizon to achieve near-optimal performance. Hence our new insight bridges the gap and interpolates between imitation learning and reinforcement learning. Motivated by the above mentioned insights, we propose Truncated HORizon Policy Search (THOR), a method that focuses on searching for policies that maximize the total reshaped reward over a finite planning horizon when the oracle is sub-optimal. We experimentally demonstrate that a gradient-based implementation of THOR can achieve superior performance compared to RL baselines and IL baselines even when the oracle is sub-optimal. |
| Tasks | Imitation Learning |
| Published | 2018-05-29 |
| URL | http://arxiv.org/abs/1805.11240v1 |
| http://arxiv.org/pdf/1805.11240v1.pdf | |
| PWC | https://paperswithcode.com/paper/truncated-horizon-policy-search-combining |
| Repo | |
| Framework | |
Data augmentation instead of explicit regularization
| Title | Data augmentation instead of explicit regularization |
| Authors | Alex Hernández-García, Peter König |
| Abstract | Modern deep artificial neural networks have achieved impressive results through models with orders of magnitude more parameters than training examples which control overfitting with the help of regularization. Regularization can be implicit, as is the case of stochastic gradient descent and parameter sharing in convolutional layers, or explicit. Explicit regularization techniques, most common forms are weight decay and dropout, have proven successful in terms of improved generalization, but they blindly reduce the effective capacity of the model, introduce sensitive hyper-parameters and require deeper and wider architectures to compensate for the reduced capacity. In contrast, data augmentation techniques exploit domain knowledge to increase the number of training examples and improve generalization without reducing the effective capacity and without introducing model-dependent parameters, since it is applied on the training data. In this paper we systematically contrast data augmentation and explicit regularization on three popular architectures and three data sets. Our results demonstrate that data augmentation alone can achieve the same performance or higher as regularized models and exhibits much higher adaptability to changes in the architecture and the amount of training data. |
| Tasks | Data Augmentation |
| Published | 2018-06-11 |
| URL | https://arxiv.org/abs/1806.03852v4 |
| https://arxiv.org/pdf/1806.03852v4.pdf | |
| PWC | https://paperswithcode.com/paper/data-augmentation-instead-of-explicit |
| Repo | |
| Framework | |
Review-Driven Multi-Label Music Style Classification by Exploiting Style Correlations
| Title | Review-Driven Multi-Label Music Style Classification by Exploiting Style Correlations |
| Authors | Guangxiang Zhao, Jingjing Xu, Qi Zeng, Xuancheng Ren |
| Abstract | This paper explores a new natural language processing task, review-driven multi-label music style classification. This task requires the system to identify multiple styles of music based on its reviews on websites. The biggest challenge lies in the complicated relations of music styles. It has brought failure to many multi-label classification methods. To tackle this problem, we propose a novel deep learning approach to automatically learn and exploit style correlations. The proposed method consists of two parts: a label-graph based neural network, and a soft training mechanism with correlation-based continuous label representation. Experimental results show that our approach achieves large improvements over the baselines on the proposed dataset. Especially, the micro F1 is improved from 53.9 to 64.5, and the one-error is reduced from 30.5 to 22.6. Furthermore, the visualized analysis shows that our approach performs well in capturing style correlations. |
| Tasks | Multi-Label Classification |
| Published | 2018-08-23 |
| URL | http://arxiv.org/abs/1808.07604v1 |
| http://arxiv.org/pdf/1808.07604v1.pdf | |
| PWC | https://paperswithcode.com/paper/review-driven-multi-label-music-style |
| Repo | |
| Framework | |
Off-Policy Evaluation and Learning from Logged Bandit Feedback: Error Reduction via Surrogate Policy
| Title | Off-Policy Evaluation and Learning from Logged Bandit Feedback: Error Reduction via Surrogate Policy |
| Authors | Yuan Xie, Boyi Liu, Qiang Liu, Zhaoran Wang, Yuan Zhou, Jian Peng |
| Abstract | When learning from a batch of logged bandit feedback, the discrepancy between the policy to be learned and the off-policy training data imposes statistical and computational challenges. Unlike classical supervised learning and online learning settings, in batch contextual bandit learning, one only has access to a collection of logged feedback from the actions taken by a historical policy, and expect to learn a policy that takes good actions in possibly unseen contexts. Such a batch learning setting is ubiquitous in online and interactive systems, such as ad platforms and recommendation systems. Existing approaches based on inverse propensity weights, such as Inverse Propensity Scoring (IPS) and Policy Optimizer for Exponential Models (POEM), enjoy unbiasedness but often suffer from large mean squared error. In this work, we introduce a new approach named Maximum Likelihood Inverse Propensity Scoring (MLIPS) for batch learning from logged bandit feedback. Instead of using the given historical policy as the proposal in inverse propensity weights, we estimate a maximum likelihood surrogate policy based on the logged action-context pairs, and then use this surrogate policy as the proposal. We prove that MLIPS is asymptotically unbiased, and moreover, has a smaller nonasymptotic mean squared error than IPS. Such an error reduction phenomenon is somewhat surprising as the estimated surrogate policy is less accurate than the given historical policy. Results on multi-label classification problems and a large- scale ad placement dataset demonstrate the empirical effectiveness of MLIPS. Furthermore, the proposed surrogate policy technique is complementary to existing error reduction techniques, and when combined, is able to consistently boost the performance of several widely used approaches. |
| Tasks | Multi-Label Classification, Recommendation Systems |
| Published | 2018-08-01 |
| URL | http://arxiv.org/abs/1808.00232v1 |
| http://arxiv.org/pdf/1808.00232v1.pdf | |
| PWC | https://paperswithcode.com/paper/off-policy-evaluation-and-learning-from |
| Repo | |
| Framework | |
Experience Replay for Continual Learning
| Title | Experience Replay for Continual Learning |
| Authors | David Rolnick, Arun Ahuja, Jonathan Schwarz, Timothy P. Lillicrap, Greg Wayne |
| Abstract | Continual learning is the problem of learning new tasks or knowledge while protecting old knowledge and ideally generalizing from old experience to learn new tasks faster. Neural networks trained by stochastic gradient descent often degrade on old tasks when trained successively on new tasks with different data distributions. This phenomenon, referred to as catastrophic forgetting, is considered a major hurdle to learning with non-stationary data or sequences of new tasks, and prevents networks from continually accumulating knowledge and skills. We examine this issue in the context of reinforcement learning, in a setting where an agent is exposed to tasks in a sequence. Unlike most other work, we do not provide an explicit indication to the model of task boundaries, which is the most general circumstance for a learning agent exposed to continuous experience. While various methods to counteract catastrophic forgetting have recently been proposed, we explore a straightforward, general, and seemingly overlooked solution - that of using experience replay buffers for all past events - with a mixture of on- and off-policy learning, leveraging behavioral cloning. We show that this strategy can still learn new tasks quickly yet can substantially reduce catastrophic forgetting in both Atari and DMLab domains, even matching the performance of methods that require task identities. When buffer storage is constrained, we confirm that a simple mechanism for randomly discarding data allows a limited size buffer to perform almost as well as an unbounded one. |
| Tasks | Continual Learning |
| Published | 2018-11-28 |
| URL | https://arxiv.org/abs/1811.11682v2 |
| https://arxiv.org/pdf/1811.11682v2.pdf | |
| PWC | https://paperswithcode.com/paper/experience-replay-for-continual-learning |
| Repo | |
| Framework | |
Enhancing keyword correlation for event detection in social networks using SVD and k-means: Twitter case study
| Title | Enhancing keyword correlation for event detection in social networks using SVD and k-means: Twitter case study |
| Authors | Ahmad Hany Hossny, Terry Moschou, Grant Osborne, Lewis Mitchell, Nick Lothian |
| Abstract | Extracting textual features from tweets is a challenging process due to the noisy nature of the content and the weak signal of most of the words used. In this paper, we propose using singular value decomposition (SVD) with clustering to enhance the signals of the textual features in the tweets to improve the correlation with events. The proposed technique applies SVD to the time series vector for each feature to factorize the matrix of feature/day counts, in order to ensure the independence of the feature vectors. Afterwards, the k-means clustering is applied to build a look-up table that maps members of each cluster to the cluster-centroid. The lookup table is used to map each feature in the original data to the centroid of its cluster, then we calculate the sum of the term frequency vectors of all features in each cluster to the term-frequency-vector of the cluster centroid. To test the technique we calculated the correlations of the cluster centroids with the golden standard record (GSR) vector before and after summing the vectors of the cluster members to the centroid-vector. The proposed method is applied to multiple correlation techniques including the Pearson, Spearman, distance correlation and Kendal Tao. The experiments have also considered the different word forms and lengths of the features including keywords, n-grams, skip-grams and bags-of-words. The correlation results are enhanced significantly as the highest correlation scores have increased from 0.3 to 0.6, and the average correlation scores have increased from 0.3 to 0.4. |
| Tasks | Time Series |
| Published | 2018-07-25 |
| URL | http://arxiv.org/abs/1807.09561v1 |
| http://arxiv.org/pdf/1807.09561v1.pdf | |
| PWC | https://paperswithcode.com/paper/enhancing-keyword-correlation-for-event |
| Repo | |
| Framework | |
Inverse Reinforcement Learning via Nonparametric Spatio-Temporal Subgoal Modeling
| Title | Inverse Reinforcement Learning via Nonparametric Spatio-Temporal Subgoal Modeling |
| Authors | Adrian Šošić, Elmar Rueckert, Jan Peters, Abdelhak M. Zoubir, Heinz Koeppl |
| Abstract | Advances in the field of inverse reinforcement learning (IRL) have led to sophisticated inference frameworks that relax the original modeling assumption of observing an agent behavior that reflects only a single intention. Instead of learning a global behavioral model, recent IRL methods divide the demonstration data into parts, to account for the fact that different trajectories may correspond to different intentions, e.g., because they were generated by different domain experts. In this work, we go one step further: using the intuitive concept of subgoals, we build upon the premise that even a single trajectory can be explained more efficiently locally within a certain context than globally, enabling a more compact representation of the observed behavior. Based on this assumption, we build an implicit intentional model of the agent’s goals to forecast its behavior in unobserved situations. The result is an integrated Bayesian prediction framework that significantly outperforms existing IRL solutions and provides smooth policy estimates consistent with the expert’s plan. Most notably, our framework naturally handles situations where the intentions of the agent change over time and classical IRL algorithms fail. In addition, due to its probabilistic nature, the model can be straightforwardly applied in active learning scenarios to guide the demonstration process of the expert. |
| Tasks | Active Learning |
| Published | 2018-03-01 |
| URL | http://arxiv.org/abs/1803.00444v3 |
| http://arxiv.org/pdf/1803.00444v3.pdf | |
| PWC | https://paperswithcode.com/paper/inverse-reinforcement-learning-via |
| Repo | |
| Framework | |
Hybrid Bayesian Eigenobjects: Combining Linear Subspace and Deep Network Methods for 3D Robot Vision
| Title | Hybrid Bayesian Eigenobjects: Combining Linear Subspace and Deep Network Methods for 3D Robot Vision |
| Authors | Benjamin Burchfiel, George Konidaris |
| Abstract | We introduce Hybrid Bayesian Eigenobjects (HBEOs), a novel representation for 3D objects designed to allow a robot to jointly estimate the pose, class, and full 3D geometry of a novel object observed from a single viewpoint in a single practical framework. By combining both linear subspace methods and deep convolutional prediction, HBEOs efficiently learn nonlinear object representations without directly regressing into high-dimensional space. HBEOs also remove the onerous and generally impractical necessity of input data voxelization prior to inference. We experimentally evaluate the suitability of HBEOs to the challenging task of joint pose, class, and shape inference on novel objects and show that, compared to preceding work, HBEOs offer dramatically improved performance in all three tasks along with several orders of magnitude faster runtime performance. |
| Tasks | |
| Published | 2018-06-20 |
| URL | http://arxiv.org/abs/1806.07872v2 |
| http://arxiv.org/pdf/1806.07872v2.pdf | |
| PWC | https://paperswithcode.com/paper/hybrid-bayesian-eigenobjects-combining-linear |
| Repo | |
| Framework | |
