Named entity recognition (NER) is a fundamental component in many applications, such as Web Search and Voice Assistants. Although deep neural networks greatly improve the performance of NER, due to the requirement of large amounts of training data, deep neural networks can hardly scale out to many languages in an industry setting. To tackle this challenge, cross-lingual NER transfers knowledge from a rich-resource language to languages with low resources through pre-trained multilingual language models. Instead of using training data in target languages, cross-lingual NER has to rely on only training data in source languages, and optionally adds the translated training data derived from source languages. However, the existing cross-lingual NER methods do not make good use of rich unlabeled data in target languages, which is relatively easy to collect in industry applications. To address the opportunities and challenges, in this paper we describe our novel practice in Microsoft to leverage such large amounts of unlabeled data in target languages in real production settings. To effectively extract weak supervision signals from the unlabeled data, we develop a novel approach based on the ideas of semi-supervised learning and reinforcement learning. The empirical study on three benchmark data sets verifies that our approach establishes the new state-of-the-art performance with clear edges. Now, the NER techniques reported in this paper are on their way to become a fundamental component for Web ranking, Entity Pane, Answers Triggering, and Question Answering in the Microsoft Bing search engine. Moreover, our techniques will also serve as part of the Spoken Language Understanding module for a commercial voice assistant. We plan to open source the code of the prototype framework after deployment.

Lack of training data in low-resource languages presents huge challenges to sequence labeling tasks such as named entity recognition (NER) and machine reading comprehension (MRC). One major obstacle is the errors on the boundary of predicted answers. To tackle this problem, we propose CalibreNet, which predicts answers in two steps. In the first step, any existing sequence labeling method can be adopted as a base model to generate an initial answer. In the second step, CalibreNet refines the boundary of the initial answer. To tackle the challenge of lack of training data in low-resource languages, we dedicatedly develop a novel unsupervised phrase boundary recovery pre-training task to enhance the multilingual boundary detection capability of CalibreNet. Experiments on two cross-lingual benchmark datasets show that the proposed approach achieves SOTA results on zero-shot cross-lingual NER and MRC tasks.

Stochastic blockmodel (SBM) is a widely used statistical network representation model, with good interpretability, expressiveness, generalization, and flexibility, which has become prevalent and important in the field of network science over the last years. However, learning an optimal SBM for a given network is an NP-hard problem. This results in significant limitations when it comes to applications of SBMs in large-scale networks, because of the significant computational overhead of existing SBM models, as well as their learning methods. Reducing the cost of SBM learning and making it scalable for handling large-scale networks, while maintaining the good theoretical properties of SBM, remains an unresolved problem. In this work, we address this challenging task from a novel perspective of model redefinition. We propose a novel redefined SBM with Poisson distribution and its block-wise learning algorithm that can efficiently analyse large-scale networks. Extensive validation conducted on both artificial and real-world data shows that our proposed method significantly outperforms the state-of-the-art methods in terms of a reasonable trade-off between accuracy and scalability. 1

Abstract User‐generated content on the Web has become an extremely valuable source for mining and analyzing user opinions on any topic. Recent years have seen an increasing body of work investigating methods to recognize favorable and unfavorable sentiments toward specific subjects from online text. However, most of these efforts focus on English and there have been very few studies on sentiment analysis of Chinese content. This paper aims to address the unique challenges posed by Chinese sentiment analysis. We propose a rule‐based approach including two phases: (1) determining each sentence's sentiment based on word dependency, and (2) aggregating sentences to predict the document sentiment. We report the results of an experimental study comparing our approach with three machine learning‐based approaches using two sets of Chinese articles. These results illustrate the effectiveness of our proposed method and its advantages against learning‐based approaches.

This paper reports a new general framework of focused web crawling based on "relational subgroup discovery". Predicates are used explicitly to represent the relevance clues of those unvisited pages in the crawl frontier, and then first-order classification rules are induced using subgroup discovery technique. The learned relational rules with sufficient support and confidence will guide the crawling process afterwards. We present the many interesting features of our proposed first-order focused crawler, together with preliminary promising experimental results.

Link context has been exploited extensively ever since the advent of the World Wide Web, but the approach to extracting precise link context has not been fully explored and many state-of-the-art extraction methods are based on simplistic heuristics and require ad-hoc parameters. In this paper, we propose a novel two-step extraction model, which aims to systematically derive link context of quality as high as anchor text. In the macroscopic analysis step, a systematic web page structure analysis is performed to locate the content cohesive text region and potential relevant header or header like tags. In the microscopic extraction step, an English parser is used to extract the relevant sentence fragments in the text region and the nearest heading text is encompassed if the need arises. Preliminary experimental results proved our approach's effectiveness.