Paper Group ANR 62
Discourse-Aware Neural Extractive Model for Text Summarization. Efficient and Robust Reinforcement Learning with Uncertainty-based Value Expansion. Mirror Natural Evolution Strategies. Conformal Prediction based Spectral Clustering. Some Interesting Features of Memristor CNN. On the difficulty of learning and predicting the long-term dynamics of bo …
Discourse-Aware Neural Extractive Model for Text Summarization
| Title | Discourse-Aware Neural Extractive Model for Text Summarization |
| Authors | Jiacheng Xu, Zhe Gan, Yu Cheng, Jingjing Liu |
| Abstract | Recently BERT has been adopted in state-of-the-art text summarization models for document encoding. However, such BERT-based extractive models use the sentence as the minimal selection unit, which often results in redundant or uninformative phrases in the generated summaries. As BERT is pre-trained on sentence pairs, not documents, the long-range dependencies between sentences are not well captured. To address these issues, we present a graph-based discourse-aware neural summarization model - DiscoBert. By utilizing discourse segmentation to extract discourse units (instead of sentences) as candidates, DiscoBert provides a fine-grained granularity for extractive selection, which helps reduce redundancy in extracted summaries. Based on this, two discourse graphs are further proposed: ($i$) RST Graph based on RST discourse trees; and ($ii$) Coreference Graph based on coreference mentions in the document. DiscoBert first encodes the extracted discourse units with BERT, and then uses a graph convolutional network to capture the long-range dependencies among discourse units through the constructed graphs. Experimental results on two popular summarization datasets demonstrate that DiscoBert outperforms state-of-the-art methods by a significant margin. |
| Tasks | Text Summarization |
| Published | 2019-10-30 |
| URL | https://arxiv.org/abs/1910.14142v1 |
| https://arxiv.org/pdf/1910.14142v1.pdf | |
| PWC | https://paperswithcode.com/paper/discourse-aware-neural-extractive-model-for |
| Repo | |
| Framework | |
Efficient and Robust Reinforcement Learning with Uncertainty-based Value Expansion
| Title | Efficient and Robust Reinforcement Learning with Uncertainty-based Value Expansion |
| Authors | Bo Zhou, Hongsheng Zeng, Fan Wang, Yunxiang Li, Hao Tian |
| Abstract | By integrating dynamics models into model-free reinforcement learning (RL) methods, model-based value expansion (MVE) algorithms have shown a significant advantage in sample efficiency as well as value estimation. However, these methods suffer from higher function approximation errors than model-free methods in stochastic environments due to a lack of modeling the environmental randomness. As a result, their performance lags behind the best model-free algorithms in some challenging scenarios. In this paper, we propose a novel Hybrid-RL method that builds on MVE, namely the Risk Averse Value Expansion (RAVE). With imaginative rollouts generated by an ensemble of probabilistic dynamics models, we further introduce the aversion of risks by seeking the lower confidence bound of the estimation. Experiments on a range of challenging environments show that by modeling the uncertainty completely, RAVE substantially enhances the robustness of previous model-based methods, and yields state-of-the-art performance. With this technique, our solution gets the first place in NeurIPS 2019: Learn to Move. |
| Tasks | |
| Published | 2019-12-10 |
| URL | https://arxiv.org/abs/1912.05328v1 |
| https://arxiv.org/pdf/1912.05328v1.pdf | |
| PWC | https://paperswithcode.com/paper/efficient-and-robust-reinforcement-learning |
| Repo | |
| Framework | |
Mirror Natural Evolution Strategies
| Title | Mirror Natural Evolution Strategies |
| Authors | Haishan Ye, Tong Zhang |
| Abstract | Evolution Strategies such as CMA-ES (covariance matrix adaptation evolution strategy) and NES (natural evolution strategy) have been widely used in machine learning applications, where an objective function is optimized without using its derivatives. However, the convergence behaviors of these algorithms have not been carefully studied. In particular, there is no rigorous analysis for the convergence of the estimated covariance matrix, and it is unclear how does the estimated covariance matrix help the converge of the algorithm. The relationship between Evolution Strategies and derivative free optimization algorithms is also not clear. In this paper, we propose a new algorithm closely related toNES, which we call MiNES (mirror descent natural evolution strategy), for which we can establish rigorous convergence results. We show that the estimated covariance matrix of MiNES converges to the inverse of Hessian matrix of the objective function with a sublinear convergence rate. Moreover, we show that some derivative free optimization algorithms are special cases of MiNES. Our empirical studies demonstrate that MiNES is a query-efficient optimization algorithm competitive to classical algorithms including NES and CMA-ES. |
| Tasks | |
| Published | 2019-10-25 |
| URL | https://arxiv.org/abs/1910.11490v1 |
| https://arxiv.org/pdf/1910.11490v1.pdf | |
| PWC | https://paperswithcode.com/paper/mirror-natural-evolution-strategies |
| Repo | |
| Framework | |
Conformal Prediction based Spectral Clustering
| Title | Conformal Prediction based Spectral Clustering |
| Authors | Lalith Srikanth Chintalapati, Raghunatha Sarma Rachakonda |
| Abstract | Spectral Clustering(SC) is a prominent data clustering technique of recent times which has attracted much attention from researchers. It is a highly data-driven method and makes no strict assumptions on the structure of the data to be clustered. One of the central pieces of spectral clustering is the construction of an affinity matrix based on a similarity measure between data points. The way the similarity measure is defined between data points has a direct impact on the performance of the SC technique. Several attempts have been made in the direction of strengthening the pairwise similarity measure to enhance the spectral clustering. In this work, we have defined a novel affinity measure by employing the concept of non-conformity used in Conformal Prediction(CP) framework. The non-conformity based affinity captures the relationship between neighborhoods of data points and has the power to generalize the notion of contextual similarity. We have shown that this formulation of affinity measure gives good results and compares well with the state of the art methods. |
| Tasks | |
| Published | 2019-09-17 |
| URL | https://arxiv.org/abs/1909.07594v1 |
| https://arxiv.org/pdf/1909.07594v1.pdf | |
| PWC | https://paperswithcode.com/paper/conformal-prediction-based-spectral |
| Repo | |
| Framework | |
Some Interesting Features of Memristor CNN
| Title | Some Interesting Features of Memristor CNN |
| Authors | Makoto Itoh |
| Abstract | In this paper, we introduce some interesting features of a memristor CNN (Cellular Neural Network). We first show that there is the similarity between the dynamics of memristors and neurons. That is, some kind of flux-controlled memristors can not respond to the sinusoidal voltage source quickly, namely, they can not switch on' rapidly. Furthermore, these memristors have refractory period after switch on’, which means that it can not respond to further sinusoidal inputs until the flux is decreased. We next show that the memristor-coupled two-cell CNN can exhibit chaotic behavior. In this system, the memristors switch off' and on’ at irregular intervals, and the two cells are connected when either or both of the memristors switches `on’. We then propose the modified CNN model, which can hold a binary output image, even if all cells are disconnected and no signal is supplied to the cell after a certain point of time. However, the modified CNN requires power to maintain the output image, that is, it is volatile. We next propose a new memristor CNN model. It can also hold a binary output state (image), even if all cells are disconnected, and no signal is supplied to the cell, by memristor’s switching behavior. Furthermore, even if we turn off the power of the system during the computation, it can resume from the previous average output state, since the memristor CNN has functions of both short-term (volatile) memory and long-term (non-volatile) memory. The above suspend and resume feature are useful when we want to save the current state, and continue work later from the previous state. Finally, we show that the memristor CNN can exhibit interesting two-dimensional waves, if an inductor is connected to each memristor CNN cell. | |
| Tasks | |
| Published | 2019-02-14 |
| URL | http://arxiv.org/abs/1902.05167v1 |
| http://arxiv.org/pdf/1902.05167v1.pdf | |
| PWC | https://paperswithcode.com/paper/some-interesting-features-of-memristor-cnn |
| Repo | |
| Framework | |
On the difficulty of learning and predicting the long-term dynamics of bouncing objects
| Title | On the difficulty of learning and predicting the long-term dynamics of bouncing objects |
| Authors | Alberto Cenzato, Alberto Testolin, Marco Zorzi |
| Abstract | The ability to accurately predict the surrounding environment is a foundational principle of intelligence in biological and artificial agents. In recent years, a variety of approaches have been proposed for learning to predict the physical dynamics of objects interacting in a visual scene. Here we conduct a systematic empirical evaluation of several state-of-the-art unsupervised deep learning models that are considered capable of learning the spatio-temporal structure of a popular dataset composed by synthetic videos of bouncing objects. We show that most of the models indeed obtain high accuracy on the standard benchmark of predicting the next frame of a sequence, and one of them even achieves state-of-the-art performance. However, all models fall short when probed with the more challenging task of generating multiple successive frames. Our results show that the ability to perform short-term predictions does not imply that the model has captured the underlying structure and dynamics of the visual environment, thereby calling for a careful rethinking of the metrics commonly adopted for evaluating temporal models. We also investigate whether the learning outcome could be affected by the use of curriculum-based teaching. |
| Tasks | |
| Published | 2019-07-31 |
| URL | https://arxiv.org/abs/1907.13494v1 |
| https://arxiv.org/pdf/1907.13494v1.pdf | |
| PWC | https://paperswithcode.com/paper/on-the-difficulty-of-learning-and-predicting |
| Repo | |
| Framework | |
Learning to run a power network challenge for training topology controllers
| Title | Learning to run a power network challenge for training topology controllers |
| Authors | Antoine Marot, Benjamin Donnot, Camilo Romero, Luca Veyrin-Forrer, Marvin Lerousseau, Balthazar Donon, Isabelle Guyon |
| Abstract | For power grid operations, a large body of research focuses on using generation redispatching, load shedding or demand side management flexibilities. However, a less costly and potentially more flexible option would be grid topology reconfiguration, as already partially exploited by Coreso (European RSC) and RTE (French TSO) operations. Beyond previous work on branch switching, bus reconfigurations are a broader class of action and could provide some substantial benefits to route electricity and optimize the grid capacity to keep it within safety margins. Because of its non-linear and combinatorial nature, no existing optimal power flow solver can yet tackle this problem. We here propose a new framework to learn topology controllers through imitation and reinforcement learning. We present the design and the results of the first “Learning to Run a Power Network” challenge released with this framework. We finally develop a method providing performance upper-bounds (oracle), which highlights remaining unsolved challenges and suggests future directions of improvement. |
| Tasks | |
| Published | 2019-12-05 |
| URL | https://arxiv.org/abs/1912.04211v1 |
| https://arxiv.org/pdf/1912.04211v1.pdf | |
| PWC | https://paperswithcode.com/paper/learning-to-run-a-power-network-challenge-for |
| Repo | |
| Framework | |
Understanding and Quantifying Adversarial Examples Existence in Linear Classification
| Title | Understanding and Quantifying Adversarial Examples Existence in Linear Classification |
| Authors | Xupeng Shi, A. Adam Ding |
| Abstract | State-of-art deep neural networks (DNN) are vulnerable to attacks by adversarial examples: a carefully designed small perturbation to the input, that is imperceptible to human, can mislead DNN. To understand the root cause of adversarial examples, we quantify the probability of adversarial example existence for linear classifiers. Previous mathematical definition of adversarial examples only involves the overall perturbation amount, and we propose a more practical relevant definition of strong adversarial examples that separately limits the perturbation along the signal direction also. We show that linear classifiers can be made robust to strong adversarial examples attack in cases where no adversarial robust linear classifiers exist under the previous definition. The quantitative formulas are confirmed by numerical experiments using a linear support vector machine (SVM) classifier. The results suggest that designing general strong-adversarial-robust learning systems is feasible but only through incorporating human knowledge of the underlying classification problem. |
| Tasks | |
| Published | 2019-10-27 |
| URL | https://arxiv.org/abs/1910.12163v1 |
| https://arxiv.org/pdf/1910.12163v1.pdf | |
| PWC | https://paperswithcode.com/paper/understanding-and-quantifying-adversarial |
| Repo | |
| Framework | |
Recurrent Transform Learning
| Title | Recurrent Transform Learning |
| Authors | Megha Gupta, Angshul Majumdar |
| Abstract | The objective of this work is to improve the accuracy of building demand forecasting. This is a more challenging task than grid level forecasting. For the said purpose, we develop a new technique called recurrent transform learning (RTL). Two versions are proposed. The first one (RTL) is unsupervised; this is used as a feature extraction tool that is further fed into a regression model. The second formulation embeds regression into the RTL framework leading to regressing recurrent transform learning (R2TL). Forecasting experiments have been carried out on three popular publicly available datasets. Both of our proposed techniques yield results superior to the state-of-the-art like long short term memory network, echo state network and sparse coding regression. |
| Tasks | |
| Published | 2019-12-11 |
| URL | https://arxiv.org/abs/1912.05198v1 |
| https://arxiv.org/pdf/1912.05198v1.pdf | |
| PWC | https://paperswithcode.com/paper/recurrent-transform-learning |
| Repo | |
| Framework | |
Zero-shot Cross-lingual Dialogue Systems with Transferable Latent Variables
| Title | Zero-shot Cross-lingual Dialogue Systems with Transferable Latent Variables |
| Authors | Zihan Liu, Jamin Shin, Yan Xu, Genta Indra Winata, Peng Xu, Andrea Madotto, Pascale Fung |
| Abstract | Despite the surging demands for multilingual task-oriented dialog systems (e.g., Alexa, Google Home), there has been less research done in multilingual or cross-lingual scenarios. Hence, we propose a zero-shot adaptation of task-oriented dialogue system to low-resource languages. To tackle this challenge, we first use a set of very few parallel word pairs to refine the aligned cross-lingual word-level representations. We then employ a latent variable model to cope with the variance of similar sentences across different languages, which is induced by imperfect cross-lingual alignments and inherent differences in languages. Finally, the experimental results show that even though we utilize much less external resources, our model achieves better adaptation performance for natural language understanding task (i.e., the intent detection and slot filling) compared to the current state-of-the-art model in the zero-shot scenario. |
| Tasks | Intent Detection, Slot Filling |
| Published | 2019-11-11 |
| URL | https://arxiv.org/abs/1911.04081v1 |
| https://arxiv.org/pdf/1911.04081v1.pdf | |
| PWC | https://paperswithcode.com/paper/zero-shot-cross-lingual-dialogue-systems-with-1 |
| Repo | |
| Framework | |
Artificial Neural Networks and Adaptive Neuro-fuzzy Models for Prediction of Remaining Useful Life
| Title | Artificial Neural Networks and Adaptive Neuro-fuzzy Models for Prediction of Remaining Useful Life |
| Authors | Razieh Tavakoli, Mohammad Najafi, Ali Sharifara |
| Abstract | The U.S. water distribution system contains thousands of miles of pipes constructed from different materials, and of various sizes, and age. These pipes suffer from physical, environmental, structural and operational stresses, causing deterioration which eventually leads to their failure. Pipe deterioration results in increased break rates, reduced hydraulic capacity, and detrimental impacts on water quality. Therefore, it is crucial to use accurate models to forecast deterioration rates along with estimating the remaining useful life of the pipes to implement essential interference plans in order to prevent catastrophic failures. This paper discusses a computational model that forecasts the RUL of water pipes by applying Artificial Neural Networks (ANNs) as well as Adaptive Neural Fuzzy Inference System (ANFIS). These models are trained and tested acquired field data to identify the significant parameters that impact the prediction of RUL. It is concluded that, on average, with approximately 10% of wall thickness loss in existing cast iron, ductile iron, asbestos-cement, and steel water pipes, the reduction of the remaining useful life is approximately 50% |
| Tasks | |
| Published | 2019-08-28 |
| URL | https://arxiv.org/abs/1909.02115v1 |
| https://arxiv.org/pdf/1909.02115v1.pdf | |
| PWC | https://paperswithcode.com/paper/artificial-neural-networks-and-adaptive-neuro |
| Repo | |
| Framework | |
BigData Applications from Graph Analytics to Machine Learning by Aggregates in Recursion
| Title | BigData Applications from Graph Analytics to Machine Learning by Aggregates in Recursion |
| Authors | Ariyam Das, Youfu Li, Jin Wang, Mingda Li, Carlo Zaniolo |
| Abstract | In the past, the semantic issues raised by the non-monotonic nature of aggregates often prevented their use in the recursive statements of logic programs and deductive databases. However, the recently introduced notion of Pre-mappability (PreM) has shown that, in key applications of interest, aggregates can be used in recursion to optimize the perfect-model semantics of aggregate-stratified programs. Therefore we can preserve the declarative formal semantics of such programs while achieving a highly efficient operational semantics that is conducive to scalable implementations on parallel and distributed platforms. In this paper, we show that with PreM, a wide spectrum of classical algorithms of practical interest, ranging from graph analytics and dynamic programming based optimization problems to data mining and machine learning applications can be concisely expressed in declarative languages by using aggregates in recursion. Our examples are also used to show that PreM can be checked using simple techniques and templatized verification strategies. A wide range of advanced BigData applications can now be expressed declaratively in logic-based languages, including Datalog, Prolog, and even SQL, while enabling their execution with superior performance and scalability. |
| Tasks | |
| Published | 2019-09-18 |
| URL | https://arxiv.org/abs/1909.08249v1 |
| https://arxiv.org/pdf/1909.08249v1.pdf | |
| PWC | https://paperswithcode.com/paper/bigdata-applications-from-graph-analytics-to |
| Repo | |
| Framework | |
AdaLinUCB: Opportunistic Learning for Contextual Bandits
| Title | AdaLinUCB: Opportunistic Learning for Contextual Bandits |
| Authors | Xueying Guo, Xiaoxiao Wang, Xin Liu |
| Abstract | In this paper, we propose and study opportunistic contextual bandits - a special case of contextual bandits where the exploration cost varies under different environmental conditions, such as network load or return variation in recommendations. When the exploration cost is low, so is the actual regret of pulling a sub-optimal arm (e.g., trying a suboptimal recommendation). Therefore, intuitively, we could explore more when the exploration cost is relatively low and exploit more when the exploration cost is relatively high. Inspired by this intuition, for opportunistic contextual bandits with Linear payoffs, we propose an Adaptive Upper-Confidence-Bound algorithm (AdaLinUCB) to adaptively balance the exploration-exploitation trade-off for opportunistic learning. We prove that AdaLinUCB achieves O((log T)^2) problem-dependent regret upper bound, which has a smaller coefficient than that of the traditional LinUCB algorithm. Moreover, based on both synthetic and real-world dataset, we show that AdaLinUCB significantly outperforms other contextual bandit algorithms, under large exploration cost fluctuations. |
| Tasks | Multi-Armed Bandits |
| Published | 2019-02-20 |
| URL | https://arxiv.org/abs/1902.07802v2 |
| https://arxiv.org/pdf/1902.07802v2.pdf | |
| PWC | https://paperswithcode.com/paper/adalinucb-opportunistic-learning-for |
| Repo | |
| Framework | |
ALIME: Autoencoder Based Approach for Local Interpretability
| Title | ALIME: Autoencoder Based Approach for Local Interpretability |
| Authors | Sharath M. Shankaranarayana, Davor Runje |
| Abstract | Machine learning and especially deep learning have garneredtremendous popularity in recent years due to their increased performanceover other methods. The availability of large amount of data has aidedin the progress of deep learning. Nevertheless, deep learning models areopaque and often seen as black boxes. Thus, there is an inherent need tomake the models interpretable, especially so in the medical domain. Inthis work, we propose a locally interpretable method, which is inspiredby one of the recent tools that has gained a lot of interest, called localinterpretable model-agnostic explanations (LIME). LIME generates singleinstance level explanation by artificially generating a dataset aroundthe instance (by randomly sampling and using perturbations) and thentraining a local linear interpretable model. One of the major issues inLIME is the instability in the generated explanation, which is caused dueto the randomly generated dataset. Another issue in these kind of localinterpretable models is the local fidelity. We propose novel modificationsto LIME by employing an autoencoder, which serves as a better weightingfunction for the local model. We perform extensive comparisons withdifferent datasets and show that our proposed method results in bothimproved stability, as well as local fidelity. |
| Tasks | |
| Published | 2019-09-04 |
| URL | https://arxiv.org/abs/1909.02437v1 |
| https://arxiv.org/pdf/1909.02437v1.pdf | |
| PWC | https://paperswithcode.com/paper/alime-autoencoder-based-approach-for-local |
| Repo | |
| Framework | |
ShrinkML: End-to-End ASR Model Compression Using Reinforcement Learning
| Title | ShrinkML: End-to-End ASR Model Compression Using Reinforcement Learning |
| Authors | Łukasz Dudziak, Mohamed S. Abdelfattah, Ravichander Vipperla, Stefanos Laskaridis, Nicholas D. Lane |
| Abstract | End-to-end automatic speech recognition (ASR) models are increasingly large and complex to achieve the best possible accuracy. In this paper, we build an AutoML system that uses reinforcement learning (RL) to optimize the per-layer compression ratios when applied to a state-of-the-art attention based end-to-end ASR model composed of several LSTM layers. We use singular value decomposition (SVD) low-rank matrix factorization as the compression method. For our RL-based AutoML system, we focus on practical considerations such as the choice of the reward/punishment functions, the formation of an effective search space, and the creation of a representative but small data set for quick evaluation between search steps. Finally, we present accuracy results on LibriSpeech of the model compressed by our AutoML system, and we compare it to manually-compressed models. Our results show that in the absence of retraining our RL-based search is an effective and practical method to compress a production-grade ASR system. When retraining is possible, we show that our AutoML system can select better highly-compressed seed models compared to manually hand-crafted rank selection, thus allowing for more compression than previously possible. |
| Tasks | AutoML, End-To-End Speech Recognition, Model Compression, Speech Recognition |
| Published | 2019-07-08 |
| URL | https://arxiv.org/abs/1907.03540v2 |
| https://arxiv.org/pdf/1907.03540v2.pdf | |
| PWC | https://paperswithcode.com/paper/shrinkml-end-to-end-asr-model-compression |
| Repo | |
| Framework | |
