Zongcheng Ji, Tian Xia, Mei Han, Jing Xiao. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Targeting at both high efficiency and performance, we propose AlignTTS to predict the mel-spectrum in parallel. AlignTTS is based on a Feed-Forward Transformer which generates mel-spectrum from a sequence of characters, and the duration of each character is determined by a duration predictor. Instead of adopting the attention mechanism in Transformer TTS to align text to mel-spectrum, the alignment loss is presented to consider all possible alignments in training by use of dynamic programming. Experiments on the LJSpeech dataset show that our model achieves not only state-of-the-art performance which outperforms Transformer TTS by 0.03 in mean option score (MOS), but also a high efficiency which is more than 50 times faster than real-time.

This paper presents a novel approach to train GANs for discrete sequence generation without resorting to an explicit neural network as the discriminator. We show that when an alternative mini-max optimization procedure is performed for the value function where a closed form solution for the discriminator exists in the maximization step, it is equivalent to directly optimizing the Jenson-Shannon divergence (JSD) between the generator’s distribution and the empirical distribution over the training data without sampling from the generator, thus optimizing the JSD becomes computationally tractable to train the generator that generates sequences of discrete data. Extensive experiments on synthetic data and real-world tasks demonstrate significant improvements over existing methods to train GANs that generate discrete sequences.

Unsegmented text recognition is a crucial component in financial document processing systems. Financial text materials, such as receipts, transcripts, identification documents, etc. often involve critical personal information. In many circumstances, these data reside in protected servers of different institutions and must not be transferred beyond the institutional firewall. The emerging technology of Federated Learning (FL) provides a data-secure way of uniting isolated datasets in model training. Using the FL framework, text recognition models can be trained with larger collection of image samples. In previous works, federated text recognition models only deal with single-character images and alpha-numeric corpus. Such models are not competent in industrial applications, especially in Chinese text recognition problems. In this paper, we apply federated learning with a deep convolutional network to perform variable-length text string recognition with a large corpus. In our experiments, we compared two prevalent federated learning frameworks, namely, Tensorflow Federated and PySyft. Results show that federated text recognition models can achieve similar or even higher accuracy than models trained on deep learning framework. On a 5-client distributed dataset, the best character accuracy is achieved by TFF at 49.20%. Extensive experiments are also conducted to evaluate the effect of distributed data storage over the performance of trained models. TFF again achieved a maximum character precision of 54.33% with non-distributed dataset.

Multi-label classification is a learning task of predicting a set of target labels for a given example. In this paper, we propose an ensemble method for multi-label classification, which is designed to optimize a novel minimum ranking margin objective function. Moreover, a boosting-type strategy is adopted to construct an accurate multi-label ensemble from multiple weak base classifiers. Experiments on different real-world multi-label classification tasks show that better performance can be achieved compared to other well-established methods.

We present a direct multi-class boosting (DMCBoost) method for classification with the following properties: (i) instead of reducing the multi-class classification task to a set of binary classification tasks, DMCBoost directly solves the multi-class classification problem, and only requires very weak base classifiers; (ii) DMCBoost builds an ensemble classifier by directly optimizing the non-convex performance measures, including the empirical classification error and margin functions, without resorting to any upper bounds or approximations. As a non-convex optimization method, DMCBoost shows competitive or better results than state-of-the-art convex relaxation boosting methods, and it performs especially well on the noisy cases.

MIRA based tuning methods have been widely used in statistical machine translation (SMT) system with a large number of features.Since the corpus-level BLEU is not decomposable, these MIRA approaches usually define a variety of heuristic-driven sentencelevel BLEUs in their model losses.Instead, we present a new MIRA method, which employs an exact corpus-level BLEU to compute the model loss.Our method is simpler in implementation.Experiments on Chinese-to-English translation show its effectiveness over two state-of-the-art MIRA implementations.

This paper proposes a multi-objective optimization framework which supports heterogeneous information sources to improve alignment in machine translation system combination techniques.In this area, most of techniques usually utilize confusion networks (CN) as their central data structure to compact an exponential number of an potential hypotheses, and because better hypothesis alignment may benefit constructing better quality confusion networks, it is natural to add more useful information to improve alignment results.However, these information may be heterogeneous, so the widely-used Viterbi algorithm for searching the best alignment may not apply here.In the multi-objective optimization framework, each information source is viewed as an independent objective, and a new goal of improving all objectives can be searched by mature algorithms.The solutions from this framework, termed Pareto optimal solutions, are then combined to construct confusion networks.Experiments on two Chinese-to-English translation datasets show significant improvements, 0.97 and 1.06 BLEU points over a strong Indirected Hidden Markov Model-based (IHMM) system, and 4.75 and 3.53 points over the best single machine translation systems.

We present a novel learning algorithm, DirectRank, which directly and exactly optimizes ranking measures without resorting to any upper bounds or approximations. Our approach is essentially an iterative coordinate ascent method. In each iteration, we choose one coordinate and only update the corresponding parameter, with all others remaining fixed. Since the ranking measure is a stepwise function of a single parameter, we propose a novel line search algorithm that can locate the interval with the best ranking measure along this coordinate quite efficiently. In order to stabilize our system in small datasets, we construct a probabilistic framework for document-query pairs to maximize the likelihood of the objective permutation of top-$\tau$ documents. This iterative procedure ensures convergence. Furthermore, we integrate regression trees as our weak learners in order to consider the correlation between the different features. Experiments on LETOR datasets and two large datasets, Yahoo challenge data and Microsoft 30K web data, show an improvement over state-of-the-art systems.

We propose a boosting method, DirectBoost, a greedy coordinate descent algorithm that builds an ensemble classifier of weak classifiers through directly minimizing empirical classification error over labeled training examples; once the training classification error is reduced to a local coordinatewise minimum, DirectBoost runs a greedy coordinate ascent algorithm that continuously adds weak classifiers to maximize any targeted arbitrarily defined margins until reaching a local coordinatewise maximum of the margins in a certain sense. Experimental results on a collection of machine-learning benchmark datasets show that DirectBoost gives consistently better results than AdaBoost, LogitBoost, LPBoost with column generation and BrownBoost, and is noise tolerant when it maximizes an n'th order bottom sample margin.