arXiv subjects

Zhiting Hu

Publications and source records attributed to Zhiting Hu.

96 records · Page 6Linked to original sources

arXiv2016

Neural Machine Translation with Recurrent Attention Modeling

Knowing which words have been attended to in previous time steps while generating a translation is a rich source of information for predicting what words will be attended to in the future. We improve upon the attention model of Bahdanau et al. (2014) by explicitly modeling the relationship between previous and subsequent attention levels for each word using one recurrent network per input word. This architecture easily captures informative features, such as fertility and regularities in relative distortion. In experiments, we show our parameterization of attention improves translation quality.

cs.NE↗

arXiv2016

Learning Concept Taxonomies from Multi-modal Data

We study the problem of automatically building hypernym taxonomies from textual and visual data. Previous works in taxonomy induction generally ignore the increasingly prominent visual data, which encode important perceptual semantics. Instead, we propose a probabilistic model for taxonomy induction by jointly leveraging text and images. To avoid hand-crafted feature engineering, we design end-to-end features based on distributed representations of images and words. The model is discriminatively trained given a small set of existing ontologies and is capable of building full taxonomies from scratch for a collection of unseen conceptual label items with associated images. We evaluate our model and features on the WordNet hierarchies, where our system outperforms previous approaches by a large gap.

cs.CL↗

arXiv2016

Joint Embeddings of Hierarchical Categories and Entities

Due to the lack of structured knowledge applied in learning distributed representation of categories, existing work cannot incorporate category hierarchies into entity information.~We propose a framework that embeds entities and categories into a semantic space by integrating structured knowledge and taxonomy hierarchy from large knowledge bases. The framework allows to compute meaningful semantic relatedness between entities and categories.~Compared with the previous state of the art, our framework can handle both single-word concepts and multiple-word concepts with superior performance in concept categorization and semantic relatedness.

cs.CL↗

arXiv2015

Poseidon: A System Architecture for Efficient GPU-based Deep Learning on Multiple Machines

Deep learning (DL) has achieved notable successes in many machine learning tasks. A number of frameworks have been developed to expedite the process of designing and training deep neural networks (DNNs), such as Caffe, Torch and Theano. Currently they can harness multiple GPUs on a single machine, but are unable to use GPUs that are distributed across multiple machines; as even average-sized DNNs can take days to train on a single GPU with 100s of GBs to TBs of data, distributed GPUs present a prime opportunity for scaling up DL. However, the limited bandwidth available on commodity Ethernet networks presents a bottleneck to distributed GPU training, and prevents its trivial realization. To investigate how to adapt existing frameworks to efficiently support distributed GPUs, we propose Poseidon, a scalable system architecture for distributed inter-machine communication in existing DL frameworks. We integrate Poseidon with Caffe and evaluate its performance at training DNNs for object recognition. Poseidon features three key contributions that accelerate DNN training on clusters: (1) a three-level hybrid architecture that allows Poseidon to support both CPU-only and GPU-equipped clusters, (2) a distributed wait-free backpropagation (DWBP) algorithm to improve GPU utilization and to balance communication, and (3) a structure-aware communication protocol (SACP) to minimize communication overheads. We empirically show that Poseidon converges to same objectives as a single machine, and achieves state-of-art training speedup across multiple models and well-established datasets using a commodity GPU cluster of 8 nodes (e.g. 4.5x speedup on AlexNet, 4x on GoogLeNet, 4x on CIFAR-10). On the much larger ImageNet22K dataset, Poseidon with 8 nodes achieves better speedup and competitive accuracy to recent CPU-based distributed systems such as Adam and Le et al., which use 10s to 1000s of nodes.

cs.LG↗

arXiv2015

Deep Kernel Learning

We introduce scalable deep kernels, which combine the structural properties of deep learning architectures with the non-parametric flexibility of kernel methods. Specifically, we transform the inputs of a spectral mixture base kernel with a deep architecture, using local kernel interpolation, inducing points, and structure exploiting (Kronecker and Toeplitz) algebra for a scalable kernel representation. These closed-form kernels can be used as drop-in replacements for standard kernels, with benefits in expressive power and scalability. We jointly learn the properties of these kernels through the marginal likelihood of a Gaussian process. Inference and learning cost $O(n)$ for $n$ training points, and predictions cost $O(1)$ per test point. On a large and diverse collection of applications, including a dataset with 2 million examples, we show improved performance over scalable Gaussian processes with flexible kernel learning models, and stand-alone deep architectures.

cs.LG↗

arXiv2013

Community Specific Temporal Topic Discovery from Social Media

Studying temporal dynamics of topics in social media is very useful to understand online user behaviors. Most of the existing work on this subject usually monitors the global trends, ignoring variation among communities. Since users from different communities tend to have varying tastes and interests, capturing community-level temporal change can improve the understanding and management of social content. Additionally, it can further facilitate the applications such as community discovery, temporal prediction and online marketing. However, this kind of extraction becomes challenging due to the intricate interactions between community and topic, and intractable computational complexity. In this paper, we take a unified solution towards the community-level topic dynamic extraction. A probabilistic model, CosTot (Community Specific Topics-over-Time) is proposed to uncover the hidden topics and communities, as well as capture community-specific temporal dynamics. Specifically, CosTot considers text, time, and network information simultaneously, and well discovers the interactions between community and topic over time. We then discuss the approximate inference implementation to enable scalable computation of model parameters, especially for large social data. Based on this, the application layer support for multi-scale temporal analysis and community exploration is also investigated. We conduct extensive experimental studies on a large real microblog dataset, and demonstrate the superiority of proposed model on tasks of time stamp prediction, link prediction and topic perplexity.

cs.SI↗