FUNSD is one of the limited publicly available datasets for information extraction from document im-ages. The information in the FUNSD dataset is defined by text areas of four categories ("key", "value", "header", "other", and "background") and connectivity between areas as key-value relations. In-specting FUNSD, we found several inconsistency in labeling, which impeded its applicability to thekey-value extraction problem. In this report, we described some labeling issues in FUNSD and therevision we made to the dataset. We also reported our implementation of for key-value detection onFUNSD using a UNet model as baseline results and an improved UNet model with Channel-InvariantDeformable Convolution.
Large pre-trained language models such as BERT have been the driving force behind recent improvements across many NLP tasks. However, BERT is only trained to predict missing words - either behind masks or in the next sentence - and has no knowledge of lexical, syntactic or semantic information beyond what it picks up through unsupervised pre-training. We propose a novel method to explicitly inject linguistic knowledge in the form of word embeddings into any layer of a pre-trained BERT. Our performance improvements on multiple semantic similarity datasets when injecting dependency-based and counter-fitted embeddings indicate that such information is beneficial and currently missing from the original model. Our qualitative analysis shows that counter-fitted embedding injection particularly helps with cases involving synonym pairs.
In this paper, incremental decode-and-forward (IDF) and incremental selective decode-and-forward (ISDF) relaying are proposed to improve the spectral efficiency of power line communication. Contrary to the traditional decode-and-forward (DF) relaying, IDF and ISDF strategies utilize the relay only if the direct link ceases to attain a certain information rate, thereby improving the spectral efficiency. The path gain through the power line is assumed to be log-normally distributed with high distance-dependent attenuation and the additive noise is from a Bernoulli-Gaussian process. Closed-form expressions for the outage probability, and approximate closed-form expressions for the end-to-end average channel capacity and the average bit error rate for binary phase-shift keying are derived. Furthermore, a closed-form expression for the fraction of times the relay is in use is derived as a measure of the spectral efficiency. Comparative analysis of IDF and ISDF with traditional DF relaying is presented. It is shown that IDF is a specific case of ISDF and can obtain optimal spectral efficiency without compromising the outage performance. By employing power allocation to minimize the outage probability, it is realized that the power should be allocated in accordance with the inter-node distances and channel parameters.
When engineers train deep learning models, they are very much "flying blind". Commonly used approaches for real-time training diagnostics, such as monitoring the train/test loss, are limited. Assessing a network's training process solely through these performance indicators is akin to debugging software without access to internal states through a debugger. To address this, we present Cockpit, a collection of instruments that enable a closer look into the inner workings of a learning machine, and a more informative and meaningful status report for practitioners. It facilitates the identification of learning phases and failure modes, like ill-chosen hyperparameters. These instruments leverage novel higher-order information about the gradient distribution and curvature, which has only recently become efficiently accessible. We believe that such a debugging tool, which we open-source for PyTorch, represents an important step to improve troubleshooting the training process, reveal new insights, and help develop novel methods and heuristics.
A novel concept of vision-based intelligent control of robotic arms is developed here in this work. This work enables the controlling of robotic arms motion only with visual inputs, that is, controlling by showing the videos of correct movements. This work can broadly be sub-divided into two segments. The first part of this work is to develop an unsupervised vision-based method to control robotic arm in 2-D plane, and the second one is with deep CNN in the same task in 3-D plane. The first method is unsupervised, where our aim is to perform mimicking of human arm motion in real-time by a manipulator. We developed a network, namely the vision-to-motion optical network (DON), where the input should be a video stream containing hand movements of human, the the output would be out the velocity and torque information of the hand movements shown in the videos. The output information of the DON is then fed to the robotic arm by enabling it to generate motion according to the real hand videos. The method has been tested with both live-stream video feed as well as on recorded video obtained from a monocular camera even by intelligently predicting the trajectory of human hand hand when it gets occluded. This is why the mimicry of the arm incorporates some intelligence to it and becomes intelligent mimic (i-mimic). Alongside the unsupervised method another method has also been developed deploying the deep neural network technique with CNN (Convolutional Neural Network) to perform the mimicking, where labelled datasets are used for training. The same dataset, as used in the unsupervised DON-based method, is used in the deep CNN method, after manual annotations. Both the proposed methods are validated with off-line as well as with on-line video datasets in real-time. The entire methodology is validated with real-time 1-link and simulated n-link manipulators alongwith suitable comparisons.
Traditionally, the efficiency and effectiveness of search systems have both been of great interest to the information retrieval community. However, an in-depth analysis of the interaction between the response latency and users' subjective search experience in the mobile setting has been missing so far. To address this gap, we conduct a controlled study that aims to reveal how response latency affects mobile web search. Our preliminary results indicate that mobile web search users are four times more tolerant to response latency reported for desktop web search users. However, when exceeding a certain threshold of 7-10 sec, the delays have a sizeable impact and users report feeling significantly more tensed, tired, terrible, frustrated and sluggish, all which contribute to a worse subjective user experience.
Recovery a planted signal perturbed by noise is a fundamental problem in machine learning. In this work, we consider the problem of recovery a planted $k$-densest sub-hypergraph on $h$-uniform hypergraphs over $n$ nodes. This fundamental problem appears in different contexts, e.g., community detection, average case complexity, and neuroscience applications. We first observe that it can be viewed as a structural variant of tensor PCA in which the hypergraph parameters $k$ and $h$ determine the structure of the signal to be recovered when the observations are contaminated by Gaussian noise. In this work, we provide tight information-theoretic upper and lower bounds for the recovery problem, as well as the first non-trivial algorithmic bounds based on approximate message passing algorithms. The problem exhibits a typical information-to-computational-gap observed in analogous settings, that widens with increasing sparsity of the problem. Interestingly, the bounds show that the structure of the signal does have an impact on the existing bounds of tensor PCA that the unstructured planted signal does not capture.
Proximity full-text search is commonly implemented in contemporary full-text search systems. Let us assume that the search query is a list of words. It is natural to consider a document as relevant if the queried words are near each other in the document. The proximity factor is even more significant for the case where the query consists of frequently occurring words. Proximity full-text search requires the storage of information for every occurrence in documents of every word that the user can search. For every occurrence of every word in a document, we employ additional indexes to store information about nearby words, that is, the words that occur in the document at distances from the given word of less than or equal to the MaxDistance parameter. We showed in previous works that these indexes can be used to improve the average query execution time by up to 130 times for queries that consist of words occurring with high-frequency. In this paper, we consider how both the search performance and the search quality depend on the value of MaxDistance and other parameters. Well-known GOV2 text collection is used in the experiments for reproducibility of the results. We propose a new index schema after the analysis of the results of the experiments. This is a pre-print of a contribution published in Supplementary Proceedings of the XXII International Conference on Data Analytics and Management in Data Intensive Domains (DAMDID/RCDL 2020), Voronezh, Russia, October 13-16, 2020, P. 336-350, published by CEUR Workshop Proceedings. The final authenticated version is available online at: http://ceur-ws.org/Vol-2790/
We tackle human image synthesis, including human motion imitation, appearance transfer, and novel view synthesis, within a unified framework. It means that the model, once being trained, can be used to handle all these tasks. The existing task-specific methods mainly use 2D keypoints to estimate the human body structure. However, they only express the position information with no abilities to characterize the personalized shape of the person and model the limb rotations. In this paper, we propose to use a 3D body mesh recovery module to disentangle the pose and shape. It can not only model the joint location and rotation but also characterize the personalized body shape. To preserve the source information, such as texture, style, color, and face identity, we propose an Attentional Liquid Warping GAN with Attentional Liquid Warping Block (AttLWB) that propagates the source information in both image and feature spaces to the synthesized reference. Specifically, the source features are extracted by a denoising convolutional auto-encoder for characterizing the source identity well. Furthermore, our proposed method can support a more flexible warping from multiple sources. To further improve the generalization ability of the unseen source images, a one/few-shot adversarial learning is applied. In detail, it firstly trains a model in an extensive training set. Then, it finetunes the model by one/few-shot unseen image(s) in a self-supervised way to generate high-resolution (512 x 512 and 1024 x 1024) results. Also, we build a new dataset, namely iPER dataset, for the evaluation of human motion imitation, appearance transfer, and novel view synthesis. Extensive experiments demonstrate the effectiveness of our methods in terms of preserving face identity, shape consistency, and clothes details. All codes and dataset are available on https://impersonator.org/work/impersonator-plus-plus.html.
Missing data can be informative. Ignoring this information can lead to misleading conclusions when the data model does not allow information to be extracted from the missing data. We propose a co-clustering model, based on the Latent Block Model, that aims to take advantage of this nonignorable nonresponses, also known as Missing Not At Random data (MNAR). A variational expectation-maximization algorithm is derived to perform inference and a model selection criterion is presented. We assess the proposed approach on a simulation study, before using our model on the voting records from the lower house of the French Parliament, where our analysis brings out relevant groups of MPs and texts, together with a sensible interpretation of the behavior of non-voters.