arXiv subjects

Byoung-Tak Zhang

Publications and source records attributed to Byoung-Tak Zhang.

107 records · Page 6Linked to original sources

arXiv2019

GLAC Net: GLocal Attention Cascading Networks for Multi-image Cued Story Generation

The task of multi-image cued story generation, such as visual storytelling dataset (VIST) challenge, is to compose multiple coherent sentences from a given sequence of images. The main difficulty is how to generate image-specific sentences within the context of overall images. Here we propose a deep learning network model, GLAC Net, that generates visual stories by combining global-local (glocal) attention and context cascading mechanisms. The model incorporates two levels of attention, i.e., overall encoding level and image feature level, to construct image-dependent sentences. While standard attention configuration needs a large number of parameters, the GLAC Net implements them in a very simple way via hard connections from the outputs of encoders or image features onto the sentence generators. The coherency of the generated story is further improved by conveying (cascading) the information of the previous sentence to the next sentence serially. We evaluate the performance of the GLAC Net on the visual storytelling dataset (VIST) and achieve very competitive results compared to the state-of-the-art techniques. Our code and pre-trained models are available here.

cs.CL↗

arXiv2019

Data Interpolations in Deep Generative Models under Non-Simply-Connected Manifold Topology

Exploiting the deep generative model's remarkable ability of learning the data-manifold structure, some recent researches proposed a geometric data interpolation method based on the geodesic curves on the learned data-manifold. However, this interpolation method often gives poor results due to a topological difference between the model and the dataset. The model defines a family of simply-connected manifolds, whereas the dataset generally contains disconnected regions or holes that make them non-simply-connected. To compensate this difference, we propose a novel density regularizer that make the interpolation path circumvent the holes denoted by low probability density. We confirm that our method gives consistently better interpolation results from the experiments with real-world image datasets.

cs.LG↗

arXiv2019

Visualizing Semantic Structures of Sequential Data by Learning Temporal Dependencies

While conventional methods for sequential learning focus on interaction between consecutive inputs, we suggest a new method which captures composite semantic flows with variable-length dependencies. In addition, the semantic structures within given sequential data can be interpreted by visualizing temporal dependencies learned from the method. The proposed method, called Temporal Dependency Network (TDN), represents a video as a temporal graph whose node represents a frame of the video and whose edge represents the temporal dependency between two frames of a variable distance. The temporal dependency structure of semantic is discovered by learning parameterized kernels of graph convolutional methods. We evaluate the proposed method on the large-scale video dataset, Youtube-8M. By visualizing the temporal dependency structures as experimental results, we show that the suggested method can find the temporal dependency structures of video semantic.

cs.LG↗

arXiv2018

Answerer in Questioner's Mind: Information Theoretic Approach to Goal-Oriented Visual Dialog

Goal-oriented dialog has been given attention due to its numerous applications in artificial intelligence. Goal-oriented dialogue tasks occur when a questioner asks an action-oriented question and an answerer responds with the intent of letting the questioner know a correct action to take. To ask the adequate question, deep learning and reinforcement learning have been recently applied. However, these approaches struggle to find a competent recurrent neural questioner, owing to the complexity of learning a series of sentences. Motivated by theory of mind, we propose "Answerer in Questioner's Mind" (AQM), a novel information theoretic algorithm for goal-oriented dialog. With AQM, a questioner asks and infers based on an approximated probabilistic model of the answerer. The questioner figures out the answerer's intention via selecting a plausible question by explicitly calculating the information gain of the candidate intentions and possible answers to each question. We test our framework on two goal-oriented visual dialog tasks: "MNIST Counting Dialog" and "GuessWhat?!". In our experiments, AQM outperforms comparative algorithms by a large margin.

cs.CV↗

arXiv2018

Bilinear Attention Networks

Attention networks in multimodal learning provide an efficient way to utilize given visual information selectively. However, the computational cost to learn attention distributions for every pair of multimodal input channels is prohibitively expensive. To solve this problem, co-attention builds two separate attention distributions for each modality neglecting the interaction between multimodal inputs. In this paper, we propose bilinear attention networks (BAN) that find bilinear attention distributions to utilize given vision-language information seamlessly. BAN considers bilinear interactions among two groups of input channels, while low-rank bilinear pooling extracts the joint representations for each pair of channels. Furthermore, we propose a variant of multimodal residual networks to exploit eight-attention maps of the BAN efficiently. We quantitatively and qualitatively evaluate our model on visual question answering (VQA 2.0) and Flickr30k Entities datasets, showing that BAN significantly outperforms previous methods and achieves new state-of-the-arts on both datasets.

cs.CV↗

arXiv2018

Multimodal Dual Attention Memory for Video Story Question Answering

We propose a video story question-answering (QA) architecture, Multimodal Dual Attention Memory (MDAM). The key idea is to use a dual attention mechanism with late fusion. MDAM uses self-attention to learn the latent concepts in scene frames and captions. Given a question, MDAM uses the second attention over these latent concepts. Multimodal fusion is performed after the dual attention processes (late fusion). Using this processing pipeline, MDAM learns to infer a high-level vision-language joint representation from an abstraction of the full video content. We evaluate MDAM on PororoQA and MovieQA datasets which have large-scale QA annotations on cartoon videos and movies, respectively. For both datasets, MDAM achieves new state-of-the-art results with significant margins compared to the runner-up models. We confirm the best performance of the dual attention mechanism combined with late fusion by ablation studies. We also perform qualitative analysis by visualizing the inference mechanisms of MDAM.

cs.CV↗

arXiv2018

Overcoming Catastrophic Forgetting by Incremental Moment Matching

Catastrophic forgetting is a problem of neural networks that loses the information of the first task after training the second task. Here, we propose a method, i.e. incremental moment matching (IMM), to resolve this problem. IMM incrementally matches the moment of the posterior distribution of the neural network which is trained on the first and the second task, respectively. To make the search space of posterior parameter smooth, the IMM procedure is complemented by various transfer learning techniques including weight transfer, L2-norm of the old and the new parameter, and a variant of dropout with the old parameter. We analyze our approach on a variety of datasets including the MNIST, CIFAR-10, Caltech-UCSD-Birds, and Lifelog datasets. The experimental results show that IMM achieves state-of-the-art performance by balancing the information between an old and a new network.

cs.LG↗

arXiv2017

Visual Explanations from Hadamard Product in Multimodal Deep Networks

The visual explanation of learned representation of models helps to understand the fundamentals of learning. The attentional models of previous works used to visualize the attended regions over an image or text using their learned weights to confirm their intended mechanism. Kim et al. (2016) show that the Hadamard product in multimodal deep networks, which is well-known for the joint function of visual question answering tasks, implicitly performs an attentional mechanism for visual inputs. In this work, we extend their work to show that the Hadamard product in multimodal deep networks performs not only for visual inputs but also for textual inputs simultaneously using the proposed gradient-based visualization technique. The attentional effect of Hadamard product is visualized for both visual and textual inputs by analyzing the two inputs and an output of the Hadamard product with the proposed method and compared with learned attentional weights of a visual question answering model.

cs.CV↗

arXiv2017

Multi-focus Attention Network for Efficient Deep Reinforcement Learning

Deep reinforcement learning (DRL) has shown incredible performance in learning various tasks to the human level. However, unlike human perception, current DRL models connect the entire low-level sensory input to the state-action values rather than exploiting the relationship between and among entities that constitute the sensory input. Because of this difference, DRL needs vast amount of experience samples to learn. In this paper, we propose a Multi-focus Attention Network (MANet) which mimics human ability to spatially abstract the low-level sensory input into multiple entities and attend to them simultaneously. The proposed method first divides the low-level input into several segments which we refer to as partial states. After this segmentation, parallel attention layers attend to the partial states relevant to solving the task. Our model estimates state-action values using these attended partial states. In our experiments, MANet attains highest scores with significantly less experience samples. Additionally, the model shows higher performance compared to the Deep Q-network and the single attention model as benchmarks. Furthermore, we extend our model to attentive communication model for performing multi-agent cooperative tasks. In multi-agent cooperative task experiments, our model shows 20% faster learning than existing state-of-the-art model.

cs.LG↗

arXiv2017

DeepStory: Video Story QA by Deep Embedded Memory Networks

Question-answering (QA) on video contents is a significant challenge for achieving human-level intelligence as it involves both vision and language in real-world settings. Here we demonstrate the possibility of an AI agent performing video story QA by learning from a large amount of cartoon videos. We develop a video-story learning model, i.e. Deep Embedded Memory Networks (DEMN), to reconstruct stories from a joint scene-dialogue video stream using a latent embedding space of observed data. The video stories are stored in a long-term memory component. For a given question, an LSTM-based attention model uses the long-term memory to recall the best question-story-answer triplet by focusing on specific words containing key information. We trained the DEMN on a novel QA dataset of children's cartoon video series, Pororo. The dataset contains 16,066 scene-dialogue pairs of 20.5-hour videos, 27,328 fine-grained sentences for scene description, and 8,913 story-related QA pairs. Our experimental results show that the DEMN outperforms other QA models. This is mainly due to 1) the reconstruction of video stories in a scene-dialogue combined form that utilize the latent embedding and 2) attention. DEMN also achieved state-of-the-art results on the MovieQA benchmark.

cs.CV↗

arXiv2017

Hadamard Product for Low-rank Bilinear Pooling

Bilinear models provide rich representations compared with linear models. They have been applied in various visual tasks, such as object recognition, segmentation, and visual question-answering, to get state-of-the-art performances taking advantage of the expanded representations. However, bilinear representations tend to be high-dimensional, limiting the applicability to computationally complex tasks. We propose low-rank bilinear pooling using Hadamard product for an efficient attention mechanism of multimodal learning. We show that our model outperforms compact bilinear pooling in visual question-answering tasks with the state-of-the-art results on the VQA dataset, having a better parsimonious property.

cs.CV↗

arXiv2017

Micro-Objective Learning : Accelerating Deep Reinforcement Learning through the Discovery of Continuous Subgoals

Recently, reinforcement learning has been successfully applied to the logical game of Go, various Atari games, and even a 3D game, Labyrinth, though it continues to have problems in sparse reward settings. It is difficult to explore, but also difficult to exploit, a small number of successes when learning policy. To solve this issue, the subgoal and option framework have been proposed. However, discovering subgoals online is too expensive to be used to learn options in large state spaces. We propose Micro-objective learning (MOL) to solve this problem. The main idea is to estimate how important a state is while training and to give an additional reward proportional to its importance. We evaluated our algorithm in two Atari games: Montezuma's Revenge and Seaquest. With three experiments to each game, MOL significantly improved the baseline scores. Especially in Montezuma's Revenge, MOL achieved two times better results than the previous state-of-the-art model.

cs.AI↗

arXiv2016

Generating Images Part by Part with Composite Generative Adversarial Networks

Image generation remains a fundamental problem in artificial intelligence in general and deep learning in specific. The generative adversarial network (GAN) was successful in generating high quality samples of natural images. We propose a model called composite generative adversarial network, that reveals the complex structure of images with multiple generators in which each generator generates some part of the image. Those parts are combined by alpha blending process to create a new single image. It can generate, for example, background and face sequentially with two generators, after training on face dataset. Training was done in an unsupervised way without any labels about what each generator should generate. We found possibilities of learning the structure by using this generative model empirically.

cs.AI↗

arXiv2016

Ways of Conditioning Generative Adversarial Networks

The GANs are generative models whose random samples realistically reflect natural images. It also can generate samples with specific attributes by concatenating a condition vector into the input, yet research on this field is not well studied. We propose novel methods of conditioning generative adversarial networks (GANs) that achieve state-of-the-art results on MNIST and CIFAR-10. We mainly introduce two models: an information retrieving model that extracts conditional information from the samples, and a spatial bilinear pooling model that forms bilinear features derived from the spatial cross product of an image and a condition vector. These methods significantly enhance log-likelihood of test data under the conditional distributions compared to the methods of concatenation.

cs.LG↗

arXiv2016

Human Body Orientation Estimation using Convolutional Neural Network

Personal robots are expected to interact with the user by recognizing the user's face. However, in most of the service robot applications, the user needs to move himself/herself to allow the robot to see him/her face to face. To overcome such limitations, a method for estimating human body orientation is required. Previous studies used various components such as feature extractors and classification models to classify the orientation which resulted in low performance. For a more robust and accurate approach, we propose the light weight convolutional neural networks, an end to end system, for estimating human body orientation. Our body orientation estimation model achieved 81.58% and 94% accuracy with the benchmark dataset and our own dataset respectively. The proposed method can be used in a wide range of service robot applications which depend on the ability to estimate human body orientation. To show its usefulness in service robot applications, we designed a simple robot application which allows the robot to move towards the user's frontal plane. With this, we demonstrated an improved face detection rate.

cs.RO↗

arXiv2016

Multimodal Residual Learning for Visual QA

Deep neural networks continue to advance the state-of-the-art of image recognition tasks with various methods. However, applications of these methods to multimodality remain limited. We present Multimodal Residual Networks (MRN) for the multimodal residual learning of visual question-answering, which extends the idea of the deep residual learning. Unlike the deep residual learning, MRN effectively learns the joint representation from vision and language information. The main idea is to use element-wise multiplication for the joint residual mappings exploiting the residual learning of the attentional models in recent studies. Various alternative models introduced by multimodality are explored based on our study. We achieve the state-of-the-art results on the Visual QA dataset for both Open-Ended and Multiple-Choice tasks. Moreover, we introduce a novel method to visualize the attention effect of the joint representations for each learning block using back-propagation algorithm, even though the visual features are collapsed without spatial information.

cs.CV↗

arXiv2015

Dual Memory Architectures for Fast Deep Learning of Stream Data via an Online-Incremental-Transfer Strategy

The online learning of deep neural networks is an interesting problem of machine learning because, for example, major IT companies want to manage the information of the massive data uploaded on the web daily, and this technology can contribute to the next generation of lifelong learning. We aim to train deep models from new data that consists of new classes, distributions, and tasks at minimal computational cost, which we call online deep learning. Unfortunately, deep neural network learning through classical online and incremental methods does not work well in both theory and practice. In this paper, we introduce dual memory architectures for online incremental deep learning. The proposed architecture consists of deep representation learners and fast learnable shallow kernel networks, both of which synergize to track the information of new data. During the training phase, we use various online, incremental ensemble, and transfer learning techniques in order to achieve lower error of the architecture. On the MNIST, CIFAR-10, and ImageNet image recognition tasks, the proposed dual memory architectures performs much better than the classical online and incremental ensemble algorithm, and their accuracies are similar to that of the batch learner.

cs.LG↗