PyTorchKR Papers

최신 논문

총 3,318편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1804.07723공식 코드 · Papers with Code

Image Inpainting for Irregular Holes Using Partial Convolutions

Existing deep learning based image inpainting methods use a standard convolutional network over the corrupted image, using convolutional filter responses conditioned on both valid pixels as well as the substitute values in the masked holes (typically the mean value). This often leads to artifacts such as color discrepancy and blurriness. Post-processing is usually used to reduce such artifacts, but are expensive and may fail. We propose the use of partial convolutions, where the convolution is masked and renormalized to be conditioned on only valid pixels. We further include a mechanism to automatically generate an updated mask for the next layer as part of the forward pass. Our model outperforms other methods for irregular masks. We show qualitative and quantitative comparisons with other methods to validate our approach.

arXiv:1702.07463공식 코드 · Papers with Code

Sequence Modeling via Segmentations

Segmental structure is a common pattern in many types of sequences such as phrases in human languages. In this paper, we present a probabilistic model for sequences via their segmentations. The probability of a segmented sequence is calculated as the product of the probabilities of all its segments, where each segment is modeled using existing tools such as recurrent neural networks. Since the segmentation of a sequence is usually unknown in advance, we sum over all valid segmentations to obtain the final probability for the sequence. An efficient dynamic programming algorithm is developed for forward and backward computations without resorting to any approximation. We demonstrate our approach on text segmentation and speech recognition tasks. In addition to quantitative results, we also show that our approach can discover meaningful segments in their respective application contexts.

arXiv:1801.07299공식 코드 · Papers with Code

Multi-Objective De Novo Drug Design with Conditional Graph Generative Model

Recently, deep generative models have revealed itself as a promising way of performing de novo molecule design. However, previous research has focused mainly on generating SMILES strings instead of molecular graphs. Although current graph generative models are available, they are often too general and computationally expensive, which restricts their application to molecules with small sizes. In this work, a new de novo molecular design framework is proposed based on a type sequential graph generators that do not use atom level recurrent units. Compared with previous graph generative models, the proposed method is much more tuned for molecule generation and have been scaled up to cover significantly larger molecules in the ChEMBL database. It is shown that the graph-based model outperforms SMILES based models in a variety of metrics, especially in the rate of valid outputs. For the application of drug design tasks, conditional graph generative model is employed. This method offers higher flexibility compared to previous fine-tuning based approach and is suitable for generation based on multiple objectives. This approach is applied to solve several drug design problems, including the generation of compounds containing a given scaffold, generation of compounds with specific drug-likeness and synthetic accessibility requirements, as well as generating dual inhibitors against JNK3 and GSK3$\beta$. Results show high enrichment rates for outputs satisfying the given requirements.

arXiv:1804.08274

Jointly Localizing and Describing Events for Dense Video Captioning

Automatically describing a video with natural language is regarded as a fundamental challenge in computer vision. The problem nevertheless is not trivial especially when a video contains multiple events to be worthy of mention, which often happens in real videos. A valid question is how to temporally localize and then describe events, which is known as "dense video captioning." In this paper, we present a novel framework for dense video captioning that unifies the localization of temporal event proposals and sentence generation of each proposal, by jointly training them in an end-to-end manner. To combine these two worlds, we integrate a new design, namely descriptiveness regression, into a single shot detection structure to infer the descriptive complexity of each detected proposal via sentence generation. This in turn adjusts the temporal locations of each event proposal. Our model differs from existing dense video captioning methods since we propose a joint and global optimization of detection and captioning, and the framework uniquely capitalizes on an attribute-augmented video captioning architecture. Extensive experiments are conducted on ActivityNet Captions dataset and our framework shows clear improvements when compared to the state-of-the-art techniques. More remarkably, we obtain a new record: METEOR of 12.96% on ActivityNet Captions official test set.

arXiv:1804.08833

Learning Manifolds from Non-stationary Streaming Data

Streaming adaptations of manifold learning based dimensionality reduction methods, such as Isomap, are based on the assumption that a small initial batch of observations is enough for exact learning of the manifold, while remaining streaming data instances can be cheaply mapped to this manifold. However, there are no theoretical results to show that this core assumption is valid. Moreover, such methods typically assume that the underlying data distribution is stationary. Such methods are not equipped to detect, or handle, sudden changes or gradual drifts in the distribution that may occur when the data is streaming. We present theoretical results to show that the quality of a manifold asymptotically converges as the size of data increases. We then show that a Gaussian Process Regression (GPR) model, that uses a manifold-specific kernel function and is trained on an initial batch of sufficient size, can closely approximate the state-of-art streaming Isomap algorithms. The predictive variance obtained from the GPR prediction is then shown to be an effective detector of changes in the underlying data distribution. Results on several synthetic and real data sets show that the resulting algorithm can effectively learn lower dimensional representation of high dimensional data in a streaming setting, while identifying shifts in the generative distribution.

arXiv:1804.10247

Experimenting with robotic intra-logistics domains

We introduce the asprilo [1] framework to facilitate experimental studies of approaches addressing complex dynamic applications. For this purpose, we have chosen the domain of robotic intra-logistics. This domain is not only highly relevant in the context of today's fourth industrial revolution but it moreover combines a multitude of challenging issues within a single uniform framework. This includes multi-agent planning, reasoning about action, change, resources, strategies, etc. In return, asprilo allows users to study alternative solutions as regards effectiveness and scalability. Although asprilo relies on Answer Set Programming and Python, it is readily usable by any system complying with its fact-oriented interface format. This makes it attractive for benchmarking and teaching well beyond logic programming. More precisely, asprilo consists of a versatile benchmark generator, solution checker and visualizer as well as a bunch of reference encodings featuring various ASP techniques. Importantly, the visualizer's animation capabilities are indispensable for complex scenarios like intra-logistics in order to inspect valid as well as invalid solution candidates. Also, it allows for graphically editing benchmark layouts that can be used as a basis for generating benchmark suites. [1] asprilo stands for Answer Set Programming for robotic intra-logistics

arXiv:1802.06300

Exact and Robust Conformal Inference Methods for Predictive Machine Learning With Dependent Data

We extend conformal inference to general settings that allow for time series data. Our proposal is developed as a randomization method and accounts for potential serial dependence by including block structures in the permutation scheme. As a result, the proposed method retains the exact, model-free validity when the data are i.i.d. or more generally exchangeable, similar to usual conformal inference methods. When exchangeability fails, as is the case for common time series data, the proposed approach is approximately valid under weak assumptions on the conformity score.

arXiv:1804.10447

Generalized Logical Operations among Conditional Events

We generalize, by a progressive procedure, the notions of conjunction and disjunction of two conditional events to the case of $n$ conditional events. In our coherence-based approach, conjunctions and disjunctions are suitable conditional random quantities. We define the notion of negation, by verifying De Morgan's Laws. We also show that conjunction and disjunction satisfy the associative and commutative properties, and a monotonicity property. Then, we give some results on coherence of prevision assessments for some families of compounded conditionals; in particular we examine the Fr\'echet-Hoeffding bounds. Moreover, we study the reverse probabilistic inference from the conjunction $\mathcal{C}_{n+1}$ of $n+1$ conditional events to the family $\{\mathcal{C}_{n},E_{n+1}|H_{n+1}\}$. We consider the relation with the notion of quasi-conjunction and we examine in detail the coherence of the prevision assessments related with the conjunction of three conditional events. Based on conjunction, we also give a characterization of p-consistency and of p-entailment, with applications to several inference rules in probabilistic nonmonotonic reasoning. Finally, we examine some non p-valid inference rules; then, we illustrate by an example two methods which allow to suitably modify non p-valid inference rules in order to get inferences which are p-valid.

arXiv:1801.04342공식 코드 · Papers with Code

Combining Symbolic Expressions and Black-box Function Evaluations in Neural Programs

Neural programming involves training neural networks to learn programs, mathematics, or logic from data. Previous works have failed to achieve good generalization performance, especially on problems and programs with high complexity or on large domains. This is because they mostly rely either on black-box function evaluations that do not capture the structure of the program, or on detailed execution traces that are expensive to obtain, and hence the training data has poor coverage of the domain under consideration. We present a novel framework that utilizes black-box function evaluations, in conjunction with symbolic expressions that define relationships between the given functions. We employ tree LSTMs to incorporate the structure of the symbolic expression trees. We use tree encoding for numbers present in function evaluation data, based on their decimal representation. We present an evaluation benchmark for this task to demonstrate our proposed model combines symbolic reasoning and function evaluation in a fruitful manner, obtaining high accuracies in our experiments. Our framework generalizes significantly better to expressions of higher depth and is able to fill partial equations with valid completions.

arXiv:1805.00121

A Missing Information Loss function for implicit feedback datasets

Latent factor models for Recommender Systems with implicit feedback typically treat unobserved user-item interactions (i.e. missing information) as negative feedback. This is frequently done either through negative sampling (point--wise loss) or with a ranking loss function (pair-- or list--wise estimation). Since a zero preference recommendation is a valid solution for most common objective functions, regarding unknown values as actual zeros results in users having a zero preference recommendation for most of the available items. In this paper we propose a novel objective function, the \emph{Missing Information Loss} (MIL), that explicitly forbids treating unobserved user-item interactions as positive or negative feedback. We apply this loss to both traditional Matrix Factorization and user--based Denoising Autoencoder, and compare it with other established objective functions such as cross-entropy (both point- and pair-wise) or the recently proposed multinomial log-likelihood. MIL achieves competitive performance in ranking-aware metrics when applied to three datasets. Furthermore, we show that such a relevance in the recommendation is obtained while displaying popular items less frequently (up to a $20 \%$ decrease with respect to the best competing method). This debiasing from the recommendation of popular items favours the appearance of infrequent items (up to a $50 \%$ increase of long-tail recommendations), a valuable feature for Recommender Systems with a large catalogue of products.

arXiv:1804.11254공식 코드 · Papers with Code

Inherent Biases in Reference based Evaluation for Grammatical Error Correction and Text Simplification

The prevalent use of too few references for evaluating text-to-text generation is known to bias estimates of their quality ({\it low coverage bias} or LCB). This paper shows that overcoming LCB in Grammatical Error Correction (GEC) evaluation cannot be attained by re-scaling or by increasing the number of references in any feasible range, contrary to previous suggestions. This is due to the long-tailed distribution of valid corrections for a sentence. Concretely, we show that LCB incentivizes GEC systems to avoid correcting even when they can generate a valid correction. Consequently, existing systems obtain comparable or superior performance compared to humans, by making few but targeted changes to the input. Similar effects on Text Simplification further support our claims.

arXiv:1805.01388

Multi-Source Fusion Operations in Subjective Logic

The purpose of multi-source fusion is to combine information from more than two evidence sources, or subjective opinions from multiple actors. For subjective logic, a number of different fusion operators have been proposed, each matching a fusion scenario with different assumptions. However, not all of these operators are associative, and therefore multi-source fusion is not well-defined for these settings. In this paper, we address this challenge, and define multi-source fusion for weighted belief fusion (WBF) and consensus & compromise fusion (CCF). For WBF, we show the definition to be equivalent to the intuitive formulation under the bijective mapping between subjective logic and Dirichlet evidence PDFs. For CCF, since there is no independent generalization, we show that the resulting multi-source fusion produces valid opinions, and explain why our generalization is sound. For completeness, we also provide corrections to previous results for averaging and cumulative belief fusion (ABF and CBF), as well as belief constraint fusion (BCF), which is an extension of Dempster's rule. With our generalizations of fusion operators, fusing information from multiple sources is now well-defined for all different fusion types defined in subjective logic. This enables wider applicability of subjective logic in applications where multiple actors interact.

arXiv:1805.02566

Understanding Reuse, Performance, and Hardware Cost of DNN Dataflows: A Data-Centric Approach Using MAESTRO

The data partitioning and scheduling strategies used by DNN accelerators to leverage reuse and perform staging are known as dataflow, and they directly impact the performance and energy efficiency of DNN accelerator designs. An accelerator microarchitecture dictates the dataflow(s) that can be employed to execute a layer or network. Selecting an optimal dataflow for a layer shape can have a large impact on utilization and energy efficiency, but there is a lack of understanding on the choices and consequences of dataflows, and of tools and methodologies to help architects explore the co-optimization design space. In this work, we first introduce a set of data-centric directives to concisely specify the space of DNN dataflows in a compilerfriendly form. We then show how these directives can be analyzed to infer various forms of reuse and to exploit them using hardware capabilities. We codify this analysis into an analytical cost model, MAESTRO (Modeling Accelerator Efficiency via Spatio-Temporal Reuse and Occupancy), that estimates various cost-benefit tradeoffs of a dataflow including execution time and energy efficiency for a DNN model and hardware configuration. We demonstrate the use of MAESTRO to drive a hardware design space exploration (DSE) experiment, which searches across 480M designs to identify 2.5M valid designs at an average rate of 0.17M designs per second, including Pareto-optimal throughput- and energy-optimized design points.

arXiv:1805.02142

An Image dehazing approach based on the airlight field estimation

This paper proposes a scheme for single image haze removal based on the airlight field (ALF) estimation. Conventional image dehazing methods which are based on a physical model generally take the global atmospheric light as a constant. However, the constant-airlight assumption may be unsuitable for images with large sky regions, which causes unacceptable brightness imbalance and color distortion in recovery images. This paper models the atmospheric light as a field function, and presents a maximum a-priori (MAP) method for jointly estimating the airlight field, the transmission rate and the haze free image. We also introduce a valid haze-level prior for effective estimate of transmission. Evaluation on real world images shows that the proposed approach outperforms existing methods in single image dehazing, especially when the large sky region is included.

arXiv:1801.09780

Bounded Policy Synthesis for POMDPs with Safe-Reachability Objectives

Planning robust executions under uncertainty is a fundamental challenge for building autonomous robots. Partially Observable Markov Decision Processes (POMDPs) provide a standard framework for modeling uncertainty in many applications. In this work, we study POMDPs with safe-reachability objectives, which require that with a probability above some threshold, a goal state is eventually reached while keeping the probability of visiting unsafe states below some threshold. This POMDP formulation is different from the traditional POMDP models with optimality objectives and we show that in some cases, POMDPs with safe-reachability objectives can provide a better guarantee of both safety and reachability than the existing POMDP models through an example. A key algorithmic problem for POMDPs is policy synthesis, which requires reasoning over a vast space of beliefs (probability distributions). To address this challenge, we introduce the notion of a goal-constrained belief space, which only contains beliefs reachable from the initial belief under desired executions that can achieve the given safe-reachability objective. Our method compactly represents this space over a bounded horizon using symbolic constraints, and employs an incremental Satisfiability Modulo Theories (SMT) solver to efficiently search for a valid policy over it. We evaluate our method using a case study involving a partially observable robotic domain with uncertain obstacles. The results show that our method can synthesize policies over large belief spaces with a small number of SMT solver calls by focusing on the goal-constrained belief space.

arXiv:1804.03824공식 코드 · Papers with Code

Reference-less Measure of Faithfulness for Grammatical Error Correction

We propose USim, a semantic measure for Grammatical Error Correction (GEC) that measures the semantic faithfulness of the output to the source, thereby complementing existing reference-less measures (RLMs) for measuring the output's grammaticality. USim operates by comparing the semantic symbolic structure of the source and the correction, without relying on manually-curated references. Our experiments establish the validity of USim, by showing that (1) semantic annotation can be consistently applied to ungrammatical text; (2) valid corrections obtain a high USim similarity score to the source; and (3) invalid corrections obtain a lower score.

arXiv:1805.04212

Behavior Analysis of NLI Models: Uncovering the Influence of Three Factors on Robustness

Natural Language Inference is a challenging task that has received substantial attention, and state-of-the-art models now achieve impressive test set performance in the form of accuracy scores. Here, we go beyond this single evaluation metric to examine robustness to semantically-valid alterations to the input data. We identify three factors - insensitivity, polarity and unseen pairs - and compare their impact on three SNLI models under a variety of conditions. Our results demonstrate a number of strengths and weaknesses in the models' ability to generalise to new in-domain instances. In particular, while strong performance is possible on unseen hypernyms, unseen antonyms are more challenging for all the models. More generally, the models suffer from an insensitivity to certain small but semantically significant alterations, and are also often influenced by simple statistical correlations between words and training labels. Overall, we show that evaluations of NLI models can benefit from studying the influence of factors intrinsic to the models or found in the dataset used.

arXiv:1805.04185공식 코드 · Papers with Code

Deep Neural Machine Translation with Weakly-Recurrent Units

Recurrent neural networks (RNNs) have represented for years the state of the art in neural machine translation. Recently, new architectures have been proposed, which can leverage parallel computation on GPUs better than classical RNNs. Faster training and inference combined with different sequence-to-sequence modeling also lead to performance improvements. While the new models completely depart from the original recurrent architecture, we decided to investigate how to make RNNs more efficient. In this work, we propose a new recurrent NMT architecture, called Simple Recurrent NMT, built on a class of fast and weakly-recurrent units that use layer normalization and multiple attentions. Our experiments on the WMT14 English-to-German and WMT16 English-Romanian benchmarks show that our model represents a valid alternative to LSTMs, as it can achieve better results at a significantly lower computational cost.

arXiv:1805.07233

Interpretable Parallel Recurrent Neural Networks with Convolutional Attentions for Multi-Modality Activity Modeling

Multimodal features play a key role in wearable sensor-based human activity recognition (HAR). Selecting the most salient features adaptively is a promising way to maximize the effectiveness of multimodal sensor data. In this regard, we propose a "collect fully and select wisely" principle as well as an interpretable parallel recurrent model with convolutional attentions to improve the recognition performance. We first collect modality features and the relations between each pair of features to generate activity frames, and then introduce an attention mechanism to select the most prominent regions from activity frames precisely. The selected frames not only maximize the utilization of valid features but also reduce the number of features to be computed effectively. We further analyze the accuracy and interpretability of the proposed model based on extensive experiments. The results show that our model achieves competitive performance on two benchmarked datasets and works well in real life scenarios.

arXiv:1805.06549공식 코드 · Papers with Code

Defoiling Foiled Image Captions

We address the task of detecting foiled image captions, i.e. identifying whether a caption contains a word that has been deliberately replaced by a semantically similar word, thus rendering it inaccurate with respect to the image being described. Solving this problem should in principle require a fine-grained understanding of images to detect linguistically valid perturbations in captions. In such contexts, encoding sufficiently descriptive image information becomes a key challenge. In this paper, we demonstrate that it is possible to solve this task using simple, interpretable yet powerful representations based on explicit object information. Our models achieve state-of-the-art performance on a standard dataset, with scores exceeding those achieved by humans on the task. We also measure the upper-bound performance of our models using gold standard annotations. Our analysis reveals that the simpler model performs well even without image information, suggesting that the dataset contains strong linguistic bias.