July 29, 2019

2806 words 14 mins read

Paper Group AWR 167

FeUdal Networks for Hierarchical Reinforcement Learning. Predicting SMT Solver Performance for Software Verification. Paying Attention to Multi-Word Expressions in Neural Machine Translation. Low Dose CT Image Denoising Using a Generative Adversarial Network with Wasserstein Distance and Perceptual Loss. Learning Dilation Factors for Semantic Segme …

FeUdal Networks for Hierarchical Reinforcement Learning


Title	FeUdal Networks for Hierarchical Reinforcement Learning
Authors	Alexander Sasha Vezhnevets, Simon Osindero, Tom Schaul, Nicolas Heess, Max Jaderberg, David Silver, Koray Kavukcuoglu
Abstract	We introduce FeUdal Networks (FuNs): a novel architecture for hierarchical reinforcement learning. Our approach is inspired by the feudal reinforcement learning proposal of Dayan and Hinton, and gains power and efficacy by decoupling end-to-end learning across multiple levels – allowing it to utilise different resolutions of time. Our framework employs a Manager module and a Worker module. The Manager operates at a lower temporal resolution and sets abstract goals which are conveyed to and enacted by the Worker. The Worker generates primitive actions at every tick of the environment. The decoupled structure of FuN conveys several benefits – in addition to facilitating very long timescale credit assignment it also encourages the emergence of sub-policies associated with different goals set by the Manager. These properties allow FuN to dramatically outperform a strong baseline agent on tasks that involve long-term credit assignment or memorisation. We demonstrate the performance of our proposed system on a range of tasks from the ATARI suite and also from a 3D DeepMind Lab environment.
Tasks	Hierarchical Reinforcement Learning
Published	2017-03-03
URL	http://arxiv.org/abs/1703.01161v2
PDF	http://arxiv.org/pdf/1703.01161v2.pdf
PWC	https://paperswithcode.com/paper/feudal-networks-for-hierarchical
Repo	https://github.com/danielpalen/pysc2-rl-agents
Framework	tf

Predicting SMT Solver Performance for Software Verification


Title	Predicting SMT Solver Performance for Software Verification
Authors	Andrew Healy, Rosemary Monahan, James F. Power
Abstract	The Why3 IDE and verification system facilitates the use of a wide range of Satisfiability Modulo Theories (SMT) solvers through a driver-based architecture. We present Where4: a portfolio-based approach to discharge Why3 proof obligations. We use data analysis and machine learning techniques on static metrics derived from program source code. Our approach benefits software engineers by providing a single utility to delegate proof obligations to the solvers most likely to return a useful result. It does this in a time-efficient way using existing Why3 and solver installations - without requiring low-level knowledge about SMT solver operation from the user.
Tasks
Published	2017-01-30
URL	http://arxiv.org/abs/1701.08466v1
PDF	http://arxiv.org/pdf/1701.08466v1.pdf
PWC	https://paperswithcode.com/paper/predicting-smt-solver-performance-for
Repo	https://github.com/ahealy19/F-IDE-2016
Framework	none

Paying Attention to Multi-Word Expressions in Neural Machine Translation


Title	Paying Attention to Multi-Word Expressions in Neural Machine Translation
Authors	Matīss Rikters, Ondřej Bojar
Abstract	Processing of multi-word expressions (MWEs) is a known problem for any natural language processing task. Even neural machine translation (NMT) struggles to overcome it. This paper presents results of experiments on investigating NMT attention allocation to the MWEs and improving automated translation of sentences that contain MWEs in English->Latvian and English->Czech NMT systems. Two improvement strategies were explored -(1) bilingual pairs of automatically extracted MWE candidates were added to the parallel corpus used to train the NMT system, and (2) full sentences containing the automatically extracted MWE candidates were added to the parallel corpus. Both approaches allowed to increase automated evaluation results. The best result - 0.99 BLEU point increase - has been reached with the first approach, while with the second approach minimal improvements achieved. We also provide open-source software and tools used for MWE extraction and alignment inspection.
Tasks	Machine Translation
Published	2017-10-17
URL	https://arxiv.org/abs/1710.06313v2
PDF	https://arxiv.org/pdf/1710.06313v2.pdf
PWC	https://paperswithcode.com/paper/paying-attention-to-multi-word-expressions-in
Repo	https://github.com/M4t1ss/MWE-Tools
Framework	none

Low Dose CT Image Denoising Using a Generative Adversarial Network with Wasserstein Distance and Perceptual Loss


Title	Low Dose CT Image Denoising Using a Generative Adversarial Network with Wasserstein Distance and Perceptual Loss
Authors	Qingsong Yang, Pingkun Yan, Yanbo Zhang, Hengyong Yu, Yongyi Shi, Xuanqin Mou, Mannudeep K. Kalra, Ge Wang
Abstract	In this paper, we introduce a new CT image denoising method based on the generative adversarial network (GAN) with Wasserstein distance and perceptual similarity. The Wasserstein distance is a key concept of the optimal transform theory, and promises to improve the performance of the GAN. The perceptual loss compares the perceptual features of a denoised output against those of the ground truth in an established feature space, while the GAN helps migrate the data noise distribution from strong to weak. Therefore, our proposed method transfers our knowledge of visual perception to the image denoising task, is capable of not only reducing the image noise level but also keeping the critical information at the same time. Promising results have been obtained in our experiments with clinical CT images.
Tasks	Denoising, Image Denoising
Published	2017-08-03
URL	http://arxiv.org/abs/1708.00961v2
PDF	http://arxiv.org/pdf/1708.00961v2.pdf
PWC	https://paperswithcode.com/paper/low-dose-ct-image-denoising-using-a
Repo	https://github.com/daintlab/ct-denoising
Framework	tf

Learning Dilation Factors for Semantic Segmentation of Street Scenes


Title	Learning Dilation Factors for Semantic Segmentation of Street Scenes
Authors	Yang He, Margret Keuper, Bernt Schiele, Mario Fritz
Abstract	Contextual information is crucial for semantic segmentation. However, finding the optimal trade-off between keeping desired fine details and at the same time providing sufficiently large receptive fields is non trivial. This is even more so, when objects or classes present in an image significantly vary in size. Dilated convolutions have proven valuable for semantic segmentation, because they allow to increase the size of the receptive field without sacrificing image resolution. However, in current state-of-the-art methods, dilation parameters are hand-tuned and fixed. In this paper, we present an approach for learning dilation parameters adaptively per channel, consistently improving semantic segmentation results on street-scene datasets like Cityscapes and Camvid.
Tasks	Semantic Segmentation
Published	2017-09-06
URL	http://arxiv.org/abs/1709.01956v1
PDF	http://arxiv.org/pdf/1709.01956v1.pdf
PWC	https://paperswithcode.com/paper/learning-dilation-factors-for-semantic
Repo	https://github.com/SSAW14/LearnableDilationNetwork
Framework	none

Towards Building a Knowledge Base of Monetary Transactions from a News Collection


Title	Towards Building a Knowledge Base of Monetary Transactions from a News Collection
Authors	Jan R. Benetka, Krisztian Balog, Kjetil Nørvåg
Abstract	We address the problem of extracting structured representations of economic events from a large corpus of news articles, using a combination of natural language processing and machine learning techniques. The developed techniques allow for semi-automatic population of a financial knowledge base, which, in turn, may be used to support a range of data mining and exploration tasks. The key challenge we face in this domain is that the same event is often reported multiple times, with varying correctness of details. We address this challenge by first collecting all information pertinent to a given event from the entire corpus, then considering all possible representations of the event, and finally, using a supervised learning method, to rank these representations by the associated confidence scores. A main innovative element of our approach is that it jointly extracts and stores all attributes of the event as a single representation (quintuple). Using a purpose-built test set we demonstrate that our supervised learning approach can achieve 25% improvement in F1-score over baseline methods that consider the earliest, the latest or the most frequent reporting of the event.
Tasks
Published	2017-09-18
URL	http://arxiv.org/abs/1709.05743v1
PDF	http://arxiv.org/pdf/1709.05743v1.pdf
PWC	https://paperswithcode.com/paper/towards-building-a-knowledge-base-of-monetary
Repo	https://github.com/benetka/kbmt
Framework	none

CLTune: A Generic Auto-Tuner for OpenCL Kernels


Title	CLTune: A Generic Auto-Tuner for OpenCL Kernels
Authors	Cedric Nugteren, Valeriu Codreanu
Abstract	This work presents CLTune, an auto-tuner for OpenCL kernels. It evaluates and tunes kernel performance of a generic, user-defined search space of possible parameter-value combinations. Example parameters include the OpenCL workgroup size, vector data-types, tile sizes, and loop unrolling factors. CLTune can be used in the following scenarios: 1) when there are too many tunable parameters to explore manually, 2) when performance portability across OpenCL devices is desired, or 3) when the optimal parameters change based on input argument values (e.g. matrix dimensions). The auto-tuner is generic, easy to use, open-source, and supports multiple search strategies including simulated annealing and particle swarm optimisation. CLTune is evaluated on two GPU case-studies inspired by the recent successes in deep learning: 2D convolution and matrix-multiplication (GEMM). For 2D convolution, we demonstrate the need for auto-tuning by optimizing for different filter sizes, achieving performance on-par or better than the state-of-the-art. For matrix-multiplication, we use CLTune to explore a parameter space of more than two-hundred thousand configurations, we show the need for device-specific tuning, and outperform the clBLAS library on NVIDIA, AMD and Intel GPUs.
Tasks
Published	2017-03-19
URL	http://arxiv.org/abs/1703.06503v1
PDF	http://arxiv.org/pdf/1703.06503v1.pdf
PWC	https://paperswithcode.com/paper/cltune-a-generic-auto-tuner-for-opencl
Repo	https://github.com/CNugteren/CLBlast
Framework	none

Unrestricted Facial Geometry Reconstruction Using Image-to-Image Translation


Title	Unrestricted Facial Geometry Reconstruction Using Image-to-Image Translation
Authors	Matan Sela, Elad Richardson, Ron Kimmel
Abstract	It has been recently shown that neural networks can recover the geometric structure of a face from a single given image. A common denominator of most existing face geometry reconstruction methods is the restriction of the solution space to some low-dimensional subspace. While such a model significantly simplifies the reconstruction problem, it is inherently limited in its expressiveness. As an alternative, we propose an Image-to-Image translation network that jointly maps the input image to a depth image and a facial correspondence map. This explicit pixel-based mapping can then be utilized to provide high quality reconstructions of diverse faces under extreme expressions, using a purely geometric refinement process. In the spirit of recent approaches, the network is trained only with synthetic data, and is then evaluated on in-the-wild facial images. Both qualitative and quantitative analyses demonstrate the accuracy and the robustness of our approach.
Tasks	3D Face Reconstruction, Image-to-Image Translation
Published	2017-03-29
URL	http://arxiv.org/abs/1703.10131v2
PDF	http://arxiv.org/pdf/1703.10131v2.pdf
PWC	https://paperswithcode.com/paper/unrestricted-facial-geometry-reconstruction
Repo	https://github.com/matansel/pix2vertex
Framework	torch

Don’t Decay the Learning Rate, Increase the Batch Size


Title	Don’t Decay the Learning Rate, Increase the Batch Size
Authors	Samuel L. Smith, Pieter-Jan Kindermans, Chris Ying, Quoc V. Le
Abstract	It is common practice to decay the learning rate. Here we show one can usually obtain the same learning curve on both training and test sets by instead increasing the batch size during training. This procedure is successful for stochastic gradient descent (SGD), SGD with momentum, Nesterov momentum, and Adam. It reaches equivalent test accuracies after the same number of training epochs, but with fewer parameter updates, leading to greater parallelism and shorter training times. We can further reduce the number of parameter updates by increasing the learning rate $\epsilon$ and scaling the batch size $B \propto \epsilon$. Finally, one can increase the momentum coefficient $m$ and scale $B \propto 1/(1-m)$, although this tends to slightly reduce the test accuracy. Crucially, our techniques allow us to repurpose existing training schedules for large batch training with no hyper-parameter tuning. We train ResNet-50 on ImageNet to $76.1%$ validation accuracy in under 30 minutes.
Tasks
Published	2017-11-01
URL	http://arxiv.org/abs/1711.00489v2
PDF	http://arxiv.org/pdf/1711.00489v2.pdf
PWC	https://paperswithcode.com/paper/dont-decay-the-learning-rate-increase-the
Repo	https://github.com/rbkim1990/capstone-age-estimation
Framework	none

ChemGAN challenge for drug discovery: can AI reproduce natural chemical diversity?


Title	ChemGAN challenge for drug discovery: can AI reproduce natural chemical diversity?
Authors	Mostapha Benhenda
Abstract	Generating molecules with desired chemical properties is important for drug discovery. The use of generative neural networks is promising for this task. However, from visual inspection, it often appears that generated samples lack diversity. In this paper, we quantify this internal chemical diversity, and we raise the following challenge: can a nontrivial AI model reproduce natural chemical diversity for desired molecules? To illustrate this question, we consider two generative models: a Reinforcement Learning model and the recently introduced ORGAN. Both fail at this challenge. We hope this challenge will stimulate research in this direction.
Tasks	Drug Discovery
Published	2017-08-28
URL	http://arxiv.org/abs/1708.08227v3
PDF	http://arxiv.org/pdf/1708.08227v3.pdf
PWC	https://paperswithcode.com/paper/chemgan-challenge-for-drug-discovery-can-ai
Repo	https://github.com/cool21th/ai_drug_discovery
Framework	none

Constant Size Molecular Descriptors For Use With Machine Learning


Title	Constant Size Molecular Descriptors For Use With Machine Learning
Authors	Christopher R. Collins, Geoffrey J. Gordon, O. Anatole von Lilienfeld, David J. Yaron
Abstract	A set of molecular descriptors whose length is independent of molecular size is developed for machine learning models that target thermodynamic and electronic properties of molecules. These features are evaluated by monitoring performance of kernel ridge regression models on well-studied data sets of small organic molecules. The features include connectivity counts, which require only the bonding pattern of the molecule, and encoded distances, which summarize distances between both bonded and non-bonded atoms and so require the full molecular geometry. In addition to having constant size, these features summarize information regarding the local environment of atoms and bonds, such that models can take advantage of similarities resulting from the presence of similar chemical fragments across molecules. Combining these two types of features leads to models whose performance is comparable to or better than the current state of the art. The features introduced here have the advantage of leading to models that may be trained on smaller molecules and then used successfully on larger molecules.
Tasks
Published	2017-01-23
URL	http://arxiv.org/abs/1701.06649v1
PDF	http://arxiv.org/pdf/1701.06649v1.pdf
PWC	https://paperswithcode.com/paper/constant-size-molecular-descriptors-for-use
Repo	https://github.com/jsheng7/ANI1-qm7
Framework	none

Broadband DOA estimation using Convolutional neural networks trained with noise signals


Title	Broadband DOA estimation using Convolutional neural networks trained with noise signals
Authors	Soumitro Chakrabarty, Emanuël. A. P. Habets
Abstract	A convolution neural network (CNN) based classification method for broadband DOA estimation is proposed, where the phase component of the short-time Fourier transform coefficients of the received microphone signals are directly fed into the CNN and the features required for DOA estimation are learnt during training. Since only the phase component of the input is used, the CNN can be trained with synthesized noise signals, thereby making the preparation of the training data set easier compared to using speech signals. Through experimental evaluation, the ability of the proposed noise trained CNN framework to generalize to speech sources is demonstrated. In addition, the robustness of the system to noise, small perturbations in microphone positions, as well as its ability to adapt to different acoustic conditions is investigated using experiments with simulated and real data.
Tasks
Published	2017-05-02
URL	http://arxiv.org/abs/1705.00919v2
PDF	http://arxiv.org/pdf/1705.00919v2.pdf
PWC	https://paperswithcode.com/paper/broadband-doa-estimation-using-convolutional
Repo	https://github.com/Soumitro-Chakrabarty/Single-speaker-localization
Framework	none

Stochastic Generative Hashing


Title	Stochastic Generative Hashing
Authors	Bo Dai, Ruiqi Guo, Sanjiv Kumar, Niao He, Le Song
Abstract	Learning-based binary hashing has become a powerful paradigm for fast search and retrieval in massive databases. However, due to the requirement of discrete outputs for the hash functions, learning such functions is known to be very challenging. In addition, the objective functions adopted by existing hashing techniques are mostly chosen heuristically. In this paper, we propose a novel generative approach to learn hash functions through Minimum Description Length principle such that the learned hash codes maximally compress the dataset and can also be used to regenerate the inputs. We also develop an efficient learning algorithm based on the stochastic distributional gradient, which avoids the notorious difficulty caused by binary output constraints, to jointly optimize the parameters of the hash function and the associated generative model. Extensive experiments on a variety of large-scale datasets show that the proposed method achieves better retrieval results than the existing state-of-the-art methods.
Tasks
Published	2017-01-11
URL	http://arxiv.org/abs/1701.02815v2
PDF	http://arxiv.org/pdf/1701.02815v2.pdf
PWC	https://paperswithcode.com/paper/stochastic-generative-hashing
Repo	https://github.com/doubling/Stochastic_Generative_Hashing
Framework	tf

Coarse-to-Fine Lifted MAP Inference in Computer Vision


Title	Coarse-to-Fine Lifted MAP Inference in Computer Vision
Authors	Haroun Habeeb, Ankit Anand, Mausam, Parag Singla
Abstract	There is a vast body of theoretical research on lifted inference in probabilistic graphical models (PGMs). However, few demonstrations exist where lifting is applied in conjunction with top of the line applied algorithms. We pursue the applicability of lifted inference for computer vision (CV), with the insight that a globally optimal (MAP) labeling will likely have the same label for two symmetric pixels. The success of our approach lies in efficiently handling a distinct unary potential on every node (pixel), typical of CV applications. This allows us to lift the large class of algorithms that model a CV problem via PGM inference. We propose a generic template for coarse-to-fine (C2F) inference in CV, which progressively refines an initial coarsely lifted PGM for varying quality-time trade-offs. We demonstrate the performance of C2F inference by developing lifted versions of two near state-of-the-art CV algorithms for stereo vision and interactive image segmentation. We find that, against flat algorithms, the lifted versions have a much superior anytime performance, without any loss in final solution quality.
Tasks	Semantic Segmentation
Published	2017-07-22
URL	http://arxiv.org/abs/1707.07165v1
PDF	http://arxiv.org/pdf/1707.07165v1.pdf
PWC	https://paperswithcode.com/paper/coarse-to-fine-lifted-map-inference-in
Repo	https://github.com/dair-iitd/c2fi4cv
Framework	none

Photorealistic Style Transfer with Screened Poisson Equation


Title	Photorealistic Style Transfer with Screened Poisson Equation
Authors	Roey Mechrez, Eli Shechtman, Lihi Zelnik-Manor
Abstract	Recent work has shown impressive success in transferring painterly style to images. These approaches, however, fall short of photorealistic style transfer. Even when both the input and reference images are photographs, the output still exhibits distortions reminiscent of a painting. In this paper we propose an approach that takes as input a stylized image and makes it more photorealistic. It relies on the Screened Poisson Equation, maintaining the fidelity of the stylized image while constraining the gradients to those of the original input image. Our method is fast, simple, fully automatic and shows positive progress in making a stylized image photorealistic. Our results exhibit finer details and are less prone to artifacts than the state-of-the-art.
Tasks	Style Transfer
Published	2017-09-28
URL	http://arxiv.org/abs/1709.09828v1
PDF	http://arxiv.org/pdf/1709.09828v1.pdf
PWC	https://paperswithcode.com/paper/photorealistic-style-transfer-with-screened
Repo	https://github.com/roimehrez/photorealism
Framework	none