February 1, 2020

3111 words 15 mins read

Paper Group AWR 345

A Hybrid Compact Neural Architecture for Visual Place Recognition. Help, Anna! Visual Navigation with Natural Multimodal Assistance via Retrospective Curiosity-Encouraging Imitation Learning. Optimizing thermodynamic trajectories using evolutionary reinforcement learning. An Open Source and Open Hardware Deep Learning-powered Visual Navigation Engi …

A Hybrid Compact Neural Architecture for Visual Place Recognition


Title	A Hybrid Compact Neural Architecture for Visual Place Recognition
Authors	Marvin Chancán, Luis Hernandez-Nunez, Ajay Narendra, Andrew B. Barron, Michael Milford
Abstract	State-of-the-art algorithms for visual place recognition, and related visual navigation systems, can be broadly split into two categories: computer-science-oriented models including deep learning or image retrieval-based techniques with minimal biological plausibility, and neuroscience-oriented dynamical networks that model temporal properties underlying spatial navigation in the brain. In this letter, we propose a new compact and high-performing place recognition model that bridges this divide for the first time. Our approach comprises two key neural models of these categories: (1) FlyNet, a compact, sparse two-layer neural network inspired by brain architectures of fruit flies, Drosophila melanogaster, and (2) a one-dimensional continuous attractor neural network (CANN). The resulting FlyNet+CANN network incorporates the compact pattern recognition capabilities of our FlyNet model with the powerful temporal filtering capabilities of an equally compact CANN, replicating entirely in a hybrid neural implementation the functionality that yields high performance in algorithmic localization approaches like SeqSLAM. We evaluate our model, and compare it to three state-of-the-art methods, on two benchmark real-world datasets with small viewpoint variations and extreme environmental changes - achieving 87% AUC results under day to night transitions compared to 60% for Multi-Process Fusion, 46% for LoST-X and 1% for SeqSLAM, while being 6.5, 310, and 1.5 times faster, respectively.
Tasks	Image Retrieval, Visual Navigation, Visual Place Recognition
Published	2019-10-15
URL	https://arxiv.org/abs/1910.06840v3
PDF	https://arxiv.org/pdf/1910.06840v3.pdf
PWC	https://paperswithcode.com/paper/a-compact-neural-architecture-for-visual
Repo	https://github.com/mchancan/flynet
Framework	pytorch


Title	Help, Anna! Visual Navigation with Natural Multimodal Assistance via Retrospective Curiosity-Encouraging Imitation Learning
Authors	Khanh Nguyen, Hal Daumé III
Abstract	Mobile agents that can leverage help from humans can potentially accomplish more complex tasks than they could entirely on their own. We develop “Help, Anna!” (HANNA), an interactive photo-realistic simulator in which an agent fulfills object-finding tasks by requesting and interpreting natural language-and-vision assistance. An agent solving tasks in a HANNA environment can leverage simulated human assistants, called ANNA (Automatic Natural Navigation Assistants), which, upon request, provide natural language and visual instructions to direct the agent towards the goals. To address the HANNA problem, we develop a memory-augmented neural agent that hierarchically models multiple levels of decision-making, and an imitation learning algorithm that teaches the agent to avoid repeating past mistakes while simultaneously predicting its own chances of making future progress. Empirically, our approach is able to ask for help more effectively than competitive baselines and, thus, attains higher task success rate on both previously seen and previously unseen environments. We publicly release code and data at https://github.com/khanhptnk/hanna . A video demo is available at https://youtu.be/18P94aaaLKg .
Tasks	Decision Making, Imitation Learning, Visual Navigation
Published	2019-09-04
URL	https://arxiv.org/abs/1909.01871v6
PDF	https://arxiv.org/pdf/1909.01871v6.pdf
PWC	https://paperswithcode.com/paper/help-anna-visual-navigation-with-natural
Repo	https://github.com/khanhptnk/hanna
Framework	pytorch

Optimizing thermodynamic trajectories using evolutionary reinforcement learning


Title	Optimizing thermodynamic trajectories using evolutionary reinforcement learning
Authors	Chris Beeler, Uladzimir Yahorau, Rory Coles, Kyle Mills, Stephen Whitelam, Isaac Tamblyn
Abstract	Using a model heat engine we show that neural network-based reinforcement learning can identify thermodynamic trajectories of maximal efficiency. We use an evolutionary learning algorithm to evolve a population of neural networks, subject to a directive to maximize the efficiency of a trajectory composed of a set of elementary thermodynamic processes; the resulting networks learn to carry out the maximally-efficient Carnot, Stirling, or Otto cycles. Given additional irreversible processes this evolutionary scheme learns a hitherto unknown thermodynamic cycle. Our results show how the reinforcement learning strategies developed for game playing can be applied to solve physical problems conditioned upon path-extensive order parameters.
Tasks
Published	2019-03-20
URL	http://arxiv.org/abs/1903.08543v1
PDF	http://arxiv.org/pdf/1903.08543v1.pdf
PWC	https://paperswithcode.com/paper/optimizing-thermodynamic-trajectories-using
Repo	https://github.com/CLEANit/heatenginegym
Framework	none


Title	An Open Source and Open Hardware Deep Learning-powered Visual Navigation Engine for Autonomous Nano-UAVs
Authors	Daniele Palossi, Francesco Conti, Luca Benini
Abstract	Nano-size unmanned aerial vehicles (UAVs), with few centimeters of diameter and sub-10 Watts of total power budget, have so far been considered incapable of running sophisticated visual-based autonomous navigation software without external aid from base-stations, ad-hoc local positioning infrastructure, and powerful external computation servers. In this work, we present what is, to the best of our knowledge, the first 27g nano-UAV system able to run aboard an end-to-end, closed-loop visual pipeline for autonomous navigation based on a state-of-the-art deep-learning algorithm, built upon the open-source CrazyFlie 2.0 nano-quadrotor. Our visual navigation engine is enabled by the combination of an ultra-low power computing device (the GAP8 system-on-chip) with a novel methodology for the deployment of deep convolutional neural networks (CNNs). We enable onboard real-time execution of a state-of-the-art deep CNN at up to 18Hz. Field experiments demonstrate that the system’s high responsiveness prevents collisions with unexpected dynamic obstacles up to a flight speed of 1.5m/s. In addition, we also demonstrate the capability of our visual navigation engine of fully autonomous indoor navigation on a 113m previously unseen path. To share our key findings with the embedded and robotics communities and foster further developments in autonomous nano-UAVs, we publicly release all our code, datasets, and trained networks.
Tasks	Autonomous Navigation, Visual Navigation
Published	2019-05-10
URL	https://arxiv.org/abs/1905.04166v1
PDF	https://arxiv.org/pdf/1905.04166v1.pdf
PWC	https://paperswithcode.com/paper/an-open-source-and-open-hardware-deep
Repo	https://github.com/pulp-platform/pulp-dronet
Framework	none

hyppo: A Comprehensive Multivariate Hypothesis Testing Python Package


Title	hyppo: A Comprehensive Multivariate Hypothesis Testing Python Package
Authors	Sambit Panda, Satish Palaniappan, Junhao Xiong, Eric W. Bridgeford, Ronak Mehta, Cencheng Shen, Joshua T. Vogelstein
Abstract	We introduce hyppo, a unified library for performing multivariate hypothesis testing, including independence, two-sample, and k-sample testing. While many multivariate independence tests have R packages available, the interfaces are inconsistent and most are not available in Python. hyppo includes many state of the art multivariate testing procedures. The package is easy-to-use and is flexible enough to enable future extensions. The documentation and all releases are available at https://hyppo.neurodata.io.
Tasks
Published	2019-07-03
URL	https://arxiv.org/abs/1907.02088v3
PDF	https://arxiv.org/pdf/1907.02088v3.pdf
PWC	https://paperswithcode.com/paper/mgcpy-a-comprehensive-high-dimensional
Repo	https://github.com/neurodata/MGC
Framework	none

GraphACT: Accelerating GCN Training on CPU-FPGA Heterogeneous Platforms


Title	GraphACT: Accelerating GCN Training on CPU-FPGA Heterogeneous Platforms
Authors	Hanqing Zeng, Viktor Prasanna
Abstract	Graph Convolutional Networks (GCNs) have emerged as the state-of-the-art deep learning model for representation learning on graphs. It is challenging to accelerate training of GCNs, due to (1) substantial and irregular data communication to propagate information within the graph, and (2) intensive computation to propagate information along the neural network layers. To address these challenges, we design a novel accelerator for training GCNs on CPU-FPGA heterogeneous systems, by incorporating multiple algorithm-architecture co-optimizations. We first analyze the computation and communication characteristics of various GCN training algorithms, and select a subgraph-based algorithm that is well suited for hardware execution. To optimize the feature propagation within subgraphs, we propose a lightweight pre-processing step based on a graph theoretic approach. Such pre-processing performed on the CPU significantly reduces the memory access requirements and the computation to be performed on the FPGA. To accelerate the weight update in GCN layers, we propose a systolic array based design for efficient parallelization. We integrate the above optimizations into a complete hardware pipeline, and analyze its load-balance and resource utilization by accurate performance modeling. We evaluate our design on a Xilinx Alveo U200 board hosted by a 40-core Xeon server. On three large graphs, we achieve an order of magnitude training speedup with negligible accuracy loss, compared with state-of-the-art implementation on a multi-core platform.
Tasks	Representation Learning
Published	2019-12-31
URL	https://arxiv.org/abs/2001.02498v1
PDF	https://arxiv.org/pdf/2001.02498v1.pdf
PWC	https://paperswithcode.com/paper/graphact-accelerating-gcn-training-on-cpu
Repo	https://github.com/GraphSAINT/GraphACT
Framework	none

Centripetal SGD for Pruning Very Deep Convolutional Networks with Complicated Structure


Title	Centripetal SGD for Pruning Very Deep Convolutional Networks with Complicated Structure
Authors	Xiaohan Ding, Guiguang Ding, Yuchen Guo, Jungong Han
Abstract	The redundancy is widely recognized in Convolutional Neural Networks (CNNs), which enables to remove unimportant filters from convolutional layers so as to slim the network with acceptable performance drop. Inspired by the linear and combinational properties of convolution, we seek to make some filters increasingly close and eventually identical for network slimming. To this end, we propose Centripetal SGD (C-SGD), a novel optimization method, which can train several filters to collapse into a single point in the parameter hyperspace. When the training is completed, the removal of the identical filters can trim the network with NO performance loss, thus no finetuning is needed. By doing so, we have partly solved an open problem of constrained filter pruning on CNNs with complicated structure, where some layers must be pruned following others. Our experimental results on CIFAR-10 and ImageNet have justified the effectiveness of C-SGD-based filter pruning. Moreover, we have provided empirical evidences for the assumption that the redundancy in deep neural networks helps the convergence of training by showing that a redundant CNN trained using C-SGD outperforms a normally trained counterpart with the equivalent width.
Tasks
Published	2019-04-08
URL	http://arxiv.org/abs/1904.03837v1
PDF	http://arxiv.org/pdf/1904.03837v1.pdf
PWC	https://paperswithcode.com/paper/centripetal-sgd-for-pruning-very-deep
Repo	https://github.com/ShawnDing1994/Centripetal-SGD
Framework	tf

Anatomical Priors in Convolutional Networks for Unsupervised Biomedical Segmentation


Title	Anatomical Priors in Convolutional Networks for Unsupervised Biomedical Segmentation
Authors	Adrian V. Dalca, John Guttag, Mert R. Sabuncu
Abstract	We consider the problem of segmenting a biomedical image into anatomical regions of interest. We specifically address the frequent scenario where we have no paired training data that contains images and their manual segmentations. Instead, we employ unpaired segmentation images to build an anatomical prior. Critically these segmentations can be derived from imaging data from a different dataset and imaging modality than the current task. We introduce a generative probabilistic model that employs the learned prior through a convolutional neural network to compute segmentations in an unsupervised setting. We conducted an empirical analysis of the proposed approach in the context of structural brain MRI segmentation, using a multi-study dataset of more than 14,000 scans. Our results show that an anatomical prior can enable fast unsupervised segmentation which is typically not possible using standard convolutional networks. The integration of anatomical priors can facilitate CNN-based anatomical segmentation in a range of novel clinical problems, where few or no annotations are available and thus standard networks are not trainable. The code is freely available at http://github.com/adalca/neuron.
Tasks
Published	2019-03-07
URL	http://arxiv.org/abs/1903.03148v1
PDF	http://arxiv.org/pdf/1903.03148v1.pdf
PWC	https://paperswithcode.com/paper/anatomical-priors-in-convolutional-networks-1
Repo	https://github.com/adalca/neuron
Framework	tf

Compositional generalization through meta sequence-to-sequence learning


Title	Compositional generalization through meta sequence-to-sequence learning
Authors	Brenden M. Lake
Abstract	People can learn a new concept and use it compositionally, understanding how to “blicket twice” after learning how to “blicket.” In contrast, powerful sequence-to-sequence (seq2seq) neural networks fail such tests of compositionality, especially when composing new concepts together with existing concepts. In this paper, I show how memory-augmented neural networks can be trained to generalize compositionally through meta seq2seq learning. In this approach, models train on a series of seq2seq problems to acquire the compositional skills needed to solve new seq2seq problems. Meta se2seq learning solves several of the SCAN tests for compositional learning and can learn to apply implicit rules to variables.
Tasks
Published	2019-06-12
URL	https://arxiv.org/abs/1906.05381v2
PDF	https://arxiv.org/pdf/1906.05381v2.pdf
PWC	https://paperswithcode.com/paper/compositional-generalization-through-meta
Repo	https://github.com/brendenlake/meta_seq2seq
Framework	pytorch

Solving Interpretable Kernel Dimension Reduction


Title	Solving Interpretable Kernel Dimension Reduction
Authors	Chieh Wu, Jared Miller, Yale Chang, Mario Sznaier, Jennifer Dy
Abstract	Kernel dimensionality reduction (KDR) algorithms find a low dimensional representation of the original data by optimizing kernel dependency measures that are capable of capturing nonlinear relationships. The standard strategy is to first map the data into a high dimensional feature space using kernels prior to a projection onto a low dimensional space. While KDR methods can be easily solved by keeping the most dominant eigenvectors of the kernel matrix, its features are no longer easy to interpret. Alternatively, Interpretable KDR (IKDR) is different in that it projects onto a subspace \textit{before} the kernel feature mapping, therefore, the projection matrix can indicate how the original features linearly combine to form the new features. Unfortunately, the IKDR objective requires a non-convex manifold optimization that is difficult to solve and can no longer be solved by eigendecomposition. Recently, an efficient iterative spectral (eigendecomposition) method (ISM) has been proposed for this objective in the context of alternative clustering. However, ISM only provides theoretical guarantees for the Gaussian kernel. This greatly constrains ISM’s usage since any kernel method using ISM is now limited to a single kernel. This work extends the theoretical guarantees of ISM to an entire family of kernels, thereby empowering ISM to solve any kernel method of the same objective. In identifying this family, we prove that each kernel within the family has a surrogate $\Phi$ matrix and the optimal projection is formed by its most dominant eigenvectors. With this extension, we establish how a wide range of IKDR applications across different learning paradigms can be solved by ISM. To support reproducible results, the source code is made publicly available on \url{https://github.com/chieh-neu/ISM_supervised_DR}.
Tasks	Dimensionality Reduction
Published	2019-09-06
URL	https://arxiv.org/abs/1909.03093v3
PDF	https://arxiv.org/pdf/1909.03093v3.pdf
PWC	https://paperswithcode.com/paper/solving-interpretable-kernel-dimension
Repo	https://github.com/chieh-neu/ISM_supervised_DR
Framework	none

Beyond Human-Level Accuracy: Computational Challenges in Deep Learning


Title	Beyond Human-Level Accuracy: Computational Challenges in Deep Learning
Authors	Joel Hestness, Newsha Ardalani, Greg Diamos
Abstract	Deep learning (DL) research yields accuracy and product improvements from both model architecture changes and scale: larger data sets and models, and more computation. For hardware design, it is difficult to predict DL model changes. However, recent prior work shows that as dataset sizes grow, DL model accuracy and model size grow predictably. This paper leverages the prior work to project the dataset and model size growth required to advance DL accuracy beyond human-level, to frontier targets defined by machine learning experts. Datasets will need to grow $33$–$971 \times$, while models will need to grow $6.6$–$456\times$ to achieve target accuracies. We further characterize and project the computational requirements to train these applications at scale. Our characterization reveals an important segmentation of DL training challenges for recurrent neural networks (RNNs) that contrasts with prior studies of deep convolutional networks. RNNs will have comparatively moderate operational intensities and very large memory footprint requirements. In contrast to emerging accelerator designs, large-scale RNN training characteristics suggest designs with significantly larger memory capacity and on-chip caches.
Tasks
Published	2019-09-03
URL	https://arxiv.org/abs/1909.01736v1
PDF	https://arxiv.org/pdf/1909.01736v1.pdf
PWC	https://paperswithcode.com/paper/beyond-human-level-accuracy-computational
Repo	https://github.com/baidu-research/catamount
Framework	tf

HyperNOMAD: Hyperparameter optimization of deep neural networks using mesh adaptive direct search


Title	HyperNOMAD: Hyperparameter optimization of deep neural networks using mesh adaptive direct search
Authors	Dounia Lakhmiri, Sébastien Le Digabel, Christophe Tribes
Abstract	The performance of deep neural networks is highly sensitive to the choice of the hyperparameters that define the structure of the network and the learning process. When facing a new application, tuning a deep neural network is a tedious and time consuming process that is often described as a “dark art”. This explains the necessity of automating the calibration of these hyperparameters. Derivative-free optimization is a field that develops methods designed to optimize time consuming functions without relying on derivatives. This work introduces the HyperNOMAD package, an extension of the NOMAD software that applies the MADS algorithm [7] to simultaneously tune the hyperparameters responsible for both the architecture and the learning process of a deep neural network (DNN), and that allows for an important flexibility in the exploration of the search space by taking advantage of categorical variables. This new approach is tested on the MNIST and CIFAR-10 data sets and achieves results comparable to the current state of the art.
Tasks	Calibration, Hyperparameter Optimization
Published	2019-07-03
URL	https://arxiv.org/abs/1907.01698v1
PDF	https://arxiv.org/pdf/1907.01698v1.pdf
PWC	https://paperswithcode.com/paper/hypernomad-hyperparameter-optimization-of
Repo	https://github.com/DouniaLakhmiri/HyperNOMAD
Framework	pytorch

GFF: Gated Fully Fusion for Semantic Segmentation


Title	GFF: Gated Fully Fusion for Semantic Segmentation
Authors	Xiangtai Li, Houlong Zhao, Lei Han, Yunhai Tong, Kuiyuan Yang
Abstract	Semantic segmentation generates comprehensive understanding of scenes through densely predicting the category for each pixel. High-level features from Deep Convolutional Neural Networks already demonstrate their effectiveness in semantic segmentation tasks, however the coarse resolution of high-level features often leads to inferior results for small/thin objects where detailed information is important. It is natural to consider importing low level features to compensate for the lost detailed information in high-level features.Unfortunately, simply combining multi-level features suffers from the semantic gap among them. In this paper, we propose a new architecture, named Gated Fully Fusion (GFF), to selectively fuse features from multiple levels using gates in a fully connected way. Specifically, features at each level are enhanced by higher-level features with stronger semantics and lower-level features with more details, and gates are used to control the propagation of useful information which significantly reduces the noises during fusion. We achieve the state of the art results on four challenging scene parsing datasets including Cityscapes, Pascal Context, COCO-stuff and ADE20K.
Tasks	Scene Parsing, Scene Understanding, Semantic Segmentation
Published	2019-04-03
URL	https://arxiv.org/abs/1904.01803v2
PDF	https://arxiv.org/pdf/1904.01803v2.pdf
PWC	https://paperswithcode.com/paper/gff-gated-fully-fusion-for-semantic
Repo	https://github.com/nhatuan84/GFF-Gated-Fully-Fusion-for-Semantic-Segmentation
Framework	tf

Harmonic Networks with Limited Training Samples


Title	Harmonic Networks with Limited Training Samples
Authors	Matej Ulicny, Vladimir A. Krylov, Rozenn Dahyot
Abstract	Convolutional neural networks (CNNs) are very popular nowadays for image processing. CNNs allow one to learn optimal filters in a (mostly) supervised machine learning context. However this typically requires abundant labelled training data to estimate the filter parameters. Alternative strategies have been deployed for reducing the number of parameters and / or filters to be learned and thus decrease overfitting. In the context of reverting to preset filters, we propose here a computationally efficient harmonic block that uses Discrete Cosine Transform (DCT) filters in CNNs. In this work we examine the performance of harmonic networks in limited training data scenario. We validate experimentally that its performance compares well against scattering networks that use wavelets as preset filters.
Tasks	Image Classification
Published	2019-04-30
URL	http://arxiv.org/abs/1905.00135v1
PDF	http://arxiv.org/pdf/1905.00135v1.pdf
PWC	https://paperswithcode.com/paper/harmonic-networks-with-limited-training
Repo	https://github.com/matej-ulicny/harmonic-networks
Framework	pytorch

Extractive Summarization of Long Documents by Combining Global and Local Context


Title	Extractive Summarization of Long Documents by Combining Global and Local Context
Authors	Wen Xiao, Giuseppe Carenini
Abstract	In this paper, we propose a novel neural single document extractive summarization model for long documents, incorporating both the global context of the whole document and the local context within the current topic. We evaluate the model on two datasets of scientific papers, Pubmed and arXiv, where it outperforms previous work, both extractive and abstractive models, on ROUGE-1, ROUGE-2 and METEOR scores. We also show that, consistently with our goal, the benefits of our method become stronger as we apply it to longer documents. Rather surprisingly, an ablation study indicates that the benefits of our model seem to come exclusively from modeling the local context, even for the longest documents.
Tasks
Published	2019-09-17
URL	https://arxiv.org/abs/1909.08089v1
PDF	https://arxiv.org/pdf/1909.08089v1.pdf
PWC	https://paperswithcode.com/paper/extractive-summarization-of-long-documents-by
Repo	https://github.com/Wendy-Xiao/Extsumm_local_global_context
Framework	pytorch