In applications involving conversational speech, data sparsity is a limiting factor in building a better language model. We propose a simple, language-independent method to quickly harvest large amounts of data from Twitter to supplement a smaller training set that is more closely matched to the domain. The techniques lead to a significant reduction in perplexity on four low-resource languages even though the presence on Twitter of these languages is relatively small. We also find that the Twitter text is more useful for learning word classes than the in-domain text and that use of these word classes leads to further reductions in perplexity. Additionally, we introduce a method of using social and textual information to prioritize the download queue during the Twitter crawling. This maximizes the amount of useful data that can be collected, impacting both perplexity and vocabulary coverage.
University of Ottawa（渥太华大学）；Kea Text；Carleton University（卡尔顿大学）；Dalhousie University（达尔豪斯大学）
We present results of empirical studies on positive speech on Twitter. By positive speech we understand speech that works for the betterment of a given situation, in this case relations between different communities in a conflict-prone country. We worked with four Twitter data sets. Through semi-manual opinion mining, we found that positive speech accounted for < 1% of the data . In fully automated studies, we tested two approaches: unsupervised statistical analysis, and supervised text classification based on distributed word representation. We discuss benefits and challenges of those approac…
We bootstrap a state-of-the-art part-of-speech tagger to tag Italian Twitter data, in the context of the Evalita 2016 PoSTWITA shared task. We show that training the tagger on native Twitter data enriched with little amounts of specifically selected gold data and additional silver-labelled data scraped from Facebook, yields better results than using large amounts of manually annotated data from a mix of genres.
Harmful speech has various forms and it has been plaguing the social media in different ways. If we need to crackdown different degrees of hate speech and abusive behavior amongst it, the classification needs to be based on complex ramifications which needs to be defined and hold accountable for, other than racist, sexist or against some particular group and community. This paper primarily describes how we created an ontological classification of harmful speech based on degree of hateful intent, and used it to annotate twitter data accordingly. The key contribution of this paper is the new dat…
Automatic speech recognition (ASR) meets more informal and free-form input data as voice user interfaces and conversational agents such as the voice assistants such as Alexa, Google Home, etc., gain popularity. Conversational speech is both the most difficult and environmentally relevant sort of data for speech recognition. In this paper, we take a linguistic perspective, and take the French language as a case study toward disambiguation of the French homophones. Our contribution aims to provide more insight into human speech transcription accuracy in conditions to reproduce those of state-of-…
We present a neural-network based approach to classifying online hate speech in general, as well as racist and sexist speech in particular. Using pre-trained word embeddings and max/mean pooling from simple, fully-connected transformations of these embeddings, we are able to predict the occurrence of hate speech on three commonly used publicly available datasets. Our models match or outperform state of the art F1 performance on all three datasets using significantly fewer parameters and minimal feature preprocessing compared to previous methods.
For conversational large-vocabulary continuous speech recognition (LVCSR) tasks, up to about two thousand hours of audio is commonly used to train state of the art models. Collection of labeled conversational audio however, is prohibitively expensive, laborious and error-prone. Furthermore, academic corpora like Fisher English (2004) or Switchboard (1992) are inadequate to train models with sufficient accuracy in the unbounded space of conversational speech. These corpora are also timeworn due to dated acoustic telephony features and the rapid advancement of colloquial vocabulary and idiomatic…
Maharashtra Institute of Technology, Pune（浦那马哈拉施特拉理工学院）
Toxic online content has become a major issue in today's world due to an exponential increase in the use of internet by people of different cultures and educational background. Differentiating hate speech and offensive language is a key challenge in automatic detection of toxic text content. In this paper, we propose an approach to automatically classify tweets on Twitter into three classes: hateful, offensive and clean. Using Twitter dataset, we perform experiments considering n-grams as features and passing their term frequency-inverse document frequency (TFIDF) values to multiple machine le…
Existing research on fairness evaluation of document classification models mainly uses synthetic monolingual data without ground truth for author demographic attributes. In this work, we assemble and publish a multilingual Twitter corpus for the task of hate speech detection with inferred four author demographic factors: age, country, gender and race/ethnicity. The corpus covers five languages: English, Italian, Polish, Portuguese and Spanish. We evaluate the inferred demographic labels with a crowdsourcing platform, Figure Eight. To examine factors that can cause biases, we take an empirical …
National Institute of Physics, University of the Philippines, Diliman（菲律宾大学迪利曼分校国家物理研究所）
Conversations reflect the existing norms of a language. Previously, we found that utterance lengths in English fictional conversations in books and movies have shortened over a period of 200 years. In this work, we show that this shortening occurs even for a brief period of 3 years (September 2009-December 2012) using 229 million utterances from Twitter. Furthermore, the subset of geographically-tagged tweets from the United States show an inverse proportion between utterance lengths and the state-level percentage of the Black population. We argue that shortening of utterances can be explained…
University of North Texas（北德克萨斯大学）；Arizona State University（亚利桑那州立大学）
Hate speech is plaguing the cyberspace along with user-generated content. This paper investigates the role of conversational context in the annotation and detection of online hate and counter speech, where context is defined as the preceding comment in a conversation thread. We created a context-aware dataset for a 3-way classification task on Reddit comments: hate speech, counter speech, or neutral. Our analyses indicate that context is critical to identify hate and counter speech: human judgments change for most comments depending on whether we show annotators the context. A linguistic analy…
Speech acts are a way to conceptualize speech as action. This holds true for communication on any platform, including social media platforms such as Twitter. In this paper, we explored speech act recognition on Twitter by treating it as a multi-class classification problem. We created a taxonomy of six speech acts for Twitter and proposed a set of semantic and syntactic features. We trained and tested a logistic regression classifier using a data set of manually labelled tweets. Our method achieved a state-of-the-art performance with an average F1 score of more than $0.70$. We also explored cl…
Universität Hamburg（汉堡大学）；Bahir Dar University（巴赫达尔大学）；Bahir Dar Institute of Technology（巴赫达尔理工学院）
In this paper, we present an analysis of the first Ethiopic Twitter Dataset for the Amharic language targeted for recognizing abusive speech. The dataset has been collected since 2014 that is written in Fidel script. Since several languages can be written using the Fidel script, we have used the existing Amharic, Tigrinya and Ge'ez corpora to retain only the Amharic tweets. We have analyzed the tweets for abusive speech content with the following targets: Analyze the distribution and tendency of abusive speech content over time and compare the abusive speech content between a Twitter and gener…
We describe a collection of acoustic and language modeling techniques that lowered the word error rate of our English conversational telephone LVCSR system to a record 6.6% on the Switchboard subset of the Hub5 2000 evaluation testset. On the acoustic side, we use a score fusion of three strong models: recurrent nets with maxout activations, very deep convolutional nets with 3x3 kernels, and bidirectional long short-term memory nets which operate on FMLLR and i-vector features. On the language modeling side, we use an updated model "M" and hierarchical neural network LMs.
University of Oslo（奥斯陆大学）；National Library of Norway（挪威国家图书馆）；University of Bergen（卑尔根大学）；University of the Basque Country（巴斯克大学）
Norwegian Twitter data poses an interesting challenge for Natural Language Processing (NLP) tasks. These texts are difficult for models trained on standardized text in one of the two Norwegian written forms (Bokmål and Nynorsk), as they contain both the typical variation of social media text, as well as a large amount of dialectal variety. In this paper we present a novel Norwegian Twitter dataset annotated with POS-tags. We show that models trained on Universal Dependency (UD) data perform worse when evaluated against this dataset, and that models trained on Bokmål generally perform better th…
Beijing University of Posts and Telecommunications（北京邮电大学）；Renmin University of China（中国人民大学）；Brandeis University（布兰迪斯大学）
In this paper, we present a neural model for joint dropped pronoun recovery (DPR) and conversational discourse parsing (CDP) in Chinese conversational speech. We show that DPR and CDP are closely related, and a joint model benefits both tasks. We refer to our model as DiscProReco, and it first encodes the tokens in each utterance in a conversation with a directed Graph Convolutional Network (GCN). The token states for an utterance are then aggregated to produce a single state for each utterance. The utterance states are then fed into a biaffine classifier to construct a conversational discours…
We describe the latest improvements to the IBM English conversational telephone speech recognition system. Some of the techniques that were found beneficial are: maxout networks with annealed dropout rates; networks with a very large number of outputs trained on 2000 hours of data; joint modeling of partially unfolded recurrent neural networks and convolutional nets by combining the bottleneck and output layers and retraining the resulting model; and lastly, sophisticated language model rescoring with exponential and neural network LMs. These techniques result in an 8.0% word error rate on the…
University of Sheffield（谢菲尔德大学）；Southwest University（西南大学）；Cardiff University（卡迪夫大学）；DFKI GmbH（德国人工智能研究中心）；Technical University Dresden（德累斯顿工业大学）；Universitat Stuttgart（斯图加特大学）；Barcelona Supercomputing Centre（巴塞罗那超级计算中心）
In recent years, the increasing propagation of hate speech on social media and the urgent need for effective counter-measures have drawn significant investment from governments, companies, and researchers. A large number of methods have been developed for automated hate speech detection online. This aims to classify textual content into non-hate or hate speech, in which case the method may also identify the targeting characteristics (i.e., types of hate, such as race, and religion) in the hate speech. However, we notice significant difference between the performance of the two (i.e., non-hate …
We describe the 2017 version of Microsoft's conversational speech recognition system, in which we update our 2016 system with recent developments in neural-network-based acoustic and language modeling to further advance the state of the art on the Switchboard speech recognition task. The system adds a CNN-BLSTM acoustic model to the set of model architectures we combined previously, and includes character-based and dialog session aware LSTM language models in rescoring. For system combination we adopt a two-stage approach, whereby subsets of acoustic models are first combined at the senone/fra…
Govt. College of Engineering & Ceramic Technology（政府工程与陶瓷技术学院）；Indian Institute of Management Calcutta（印度管理学院加尔各答分校）
Social media platforms contain great wealth of information which provides us opportunities explore hidden patterns or unknown correlations, and understand people's satisfaction with what they are discussing. As one showcase, in this paper, we summarize the data set of Twitter messages related to recent demonetization of all Rs. 500 and Rs. 1000 notes in India and explore insights from Twitter's data. Our proposed system automatically extracts the popular latent topics in conversations regarding demonetization discussed in Twitter via the Latent Dirichlet Allocation (LDA) based topic model and …
In this paper, we describe recent performance improvements to the production Marchex speech recognition system for our spontaneous customer-to-business telephone conversations. In our previous work, we focused on in-domain language and acoustic model training. In this work we employ state-of-the-art semi-supervised lattice-free maximum mutual information (LF-MMI) training process which can supervise over full lattices from unlabeled audio. On Marchex English (ME), a modern evaluation set of conversational North American English, we observed a 3.3% (3.2% for agent, 3.6% for caller) reduction in…
Goal-oriented conversational interfaces are designed to accomplish specific tasks and typically have interactions that tend to span multiple turns adhering to a pre-defined structure and a goal. However, conventional neural language models (NLM) in Automatic Speech Recognition (ASR) systems are mostly trained sentence-wise with limited context. In this paper, we explore different ways to incorporate context into a LSTM based NLM in order to model long range dependencies and improve speech recognition. Specifically, we use context carry over across multiple turns and use lexical contextual cues…
In the past few years, there has been a significant rise in toxic and hateful content on various social media platforms. Recently Black Lives Matter movement came into the picture, causing an avalanche of user generated responses on the internet. In this paper, we have proposed a Black Lives Matter related tweet hate speech dataset TweetBLM. Our dataset comprises 9165 manually annotated tweets that target the Black Lives Matter movement. We annotated the tweets into two classes, i.e., HATE and NONHATE based on their content related to racism erupted from the movement for the black community. I…
Multimodal Fine-grained Context Interaction Graph Modeling for Conversational Speech Synthesis
Inner Mongolia University（内蒙古大学）；Center for Language and Speech Processing (CLSP)（语言与语音处理中心）；School of Artificial Intelligence, The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）人工智能学院）
We present a novel conversational-context aware end-to-end speech recognizer based on a gated neural network that incorporates conversational-context/word/speech embeddings. Unlike conventional speech recognition models, our model learns longer conversational-context information that spans across sentences and is consequently better at recognizing long conversations. Specifically, we propose to use the text-based external word and/or sentence embeddings (i.e., fastText, BERT) within an end-to-end framework, yielding a significant improvement in word error rate with better conversational-contex…
In this paper we show how we have achieved the state-of-the-art performance on the industry-standard NIST 2000 Hub5 English evaluation set. We explore densely connected LSTMs, inspired by the densely connected convolutional networks recently introduced for image classification tasks. We also propose an acoustic model adaptation scheme that simply averages the parameters of a seed neural network acoustic model and its adapted version. This method was applied with the CallHome training corpus and improved individual system performances by on average 6.1% (relative) against the CallHome portion o…
We describe a statistical approach for modeling dialogue acts in conversational speech, i.e., speech-act-like units such as Statement, Question, Backchannel, Agreement, Disagreement, and Apology. Our model detects and predicts dialogue acts based on lexical, collocational, and prosodic cues, as well as on the discourse coherence of the dialogue act sequence. The dialogue model is based on treating the discourse structure of a conversation as a hidden Markov model and the individual dialogue acts as observations emanating from the model states. Constraints on the likely sequence of dialogue act…
LSTM language models (LSTM-LMs) have been proven to be powerful and yielded significant performance improvements over count based n-gram LMs in modern speech recognition systems. Due to its infinite history states and computational load, most previous studies focus on applying LSTM-LMs in the second-pass for rescoring purpose. Recent work shows that it is feasible and computationally affordable to adopt the LSTM-LMs in the first-pass decoding within a dynamic (or tree based) decoder framework. In this work, the LSTM-LM is composed with a WFST decoder on-the-fly for the first-pass decoding. Fur…
Microsoft AI and Research（微软人工智能与研究部）；Microsoft AI（微软人工智能）；Research, Redmond, WA, USA（研究部，美国华盛顿州雷德蒙德）
Recent work in automatic recognition of conversational telephone speech (CTS) has achieved accuracy levels comparable to human transcribers, although there is some debate how to precisely quantify human performance on this task, using the NIST 2000 CTS evaluation set. This raises the question what systematic differences, if any, may be found differentiating human from machine transcription errors. In this paper we approach this question by comparing the output of our most accurate CTS recognition system to that of a standard speech transcription vendor pipeline. We find that the most frequent …
The context-dependent nature of online aggression makes annotating large collections of data extremely difficult. Previously studied datasets in abusive language detection have been insufficient in size to efficiently train deep learning models. Recently, Hate and Abusive Speech on Twitter, a dataset much greater in size and reliability, has been released. However, this dataset has not been comprehensively studied to its potential. In this paper, we conduct the first comparative study of various learning models on Hate and Abusive Speech on Twitter, and discuss the possibility of using additio…
One of the most difficult speech recognition tasks is accurate recognition of human to human communication. Advances in deep learning over the last few years have produced major speech recognition improvements on the representative Switchboard conversational corpus. Word error rates that just a few years ago were 14% have dropped to 8.0%, then 6.6% and most recently 5.8%, and are now believed to be within striking range of human performance. This then raises two issues - what IS human performance, and how far down can we still drive speech recognition error rates? A recent paper by Microsoft s…
Indian Institute of Information Technology, Kalyani（印度信息技术学院卡利亚尼分校）
Social media has seen a worrying rise in hate speech in recent times. Branching to several distinct categories of cyberbullying, gender discrimination, or racism, the combined label for such derogatory content can be classified as toxic content in general. This paper presents experimentation with a Keras wrapped lightweight BERT model to successfully identify hate speech and predict probabilistic impact score for the same to extract the hateful words within sentences. The dataset used for this task is the Hate Speech and Offensive Content Detection (HASOC 2021) data from FIRE 2021 in English. …
University of Hildesheim（希尔德斯海姆大学）；University of Antwerp（安特卫普大学）
Discussion about the social network Twitter often concerns its role in political discourse, involving the question of when an expression of opinion becomes offensive, immoral, and/or illegal, and how to deal with it. Given the growing amount of offensive communication on the internet, there is a demand for new technology that can automatically detect hate speech, to assist content moderation by humans. This comes with new challenges, such as defining exactly what is free speech and what is illegal in a specific country, and knowing exactly what the linguistic characteristics of hate speech are…
StyleBench: Evaluating Speech Language Models on Conversational Speaking Style Control
NLP Lab, School of Computer Science and Engineering, Northeastern University, Shenyang, China（计算机科学与工程学院自然语言处理实验室，东北大学，沈阳，中国）；NiuTrans Research, Shenyang, China（尼乌透研究，沈阳，中国）
Today, the vocabulary size for language models in large vocabulary speech recognition is typically several hundreds of thousands of words. While this is already sufficient in some applications, the out-of-vocabulary words are still limiting the usability in others. In agglutinative languages the vocabulary for conversational speech should include millions of word forms to cover the spelling variations due to colloquial pronunciations, in addition to the word compounding and inflections. Very large vocabularies are also needed, for example, when the recognition of rare proper names is important…
Columbia University in the City of New York（纽约哥伦比亚大学）
We propose a way to use a transformer-based language model in conversational speech recognition. Specifically, we focus on decoding efficiently in a weighted finite-state transducer framework. We showcase an approach to lattice re-scoring that allows for longer range history captured by a transfomer-based language model and takes advantage of a transformer's ability to avoid computing sequentially.
Conversational automatic speech recognition (ASR) is a task to recognize conversational speech including multiple speakers. Unlike sentence-level ASR, conversational ASR can naturally take advantages from specific characteristics of conversation, such as role preference and topical coherence. This paper proposes a conversational ASR model which explicitly learns conversation-level characteristics under the prevalent end-to-end neural framework. The highlights of the proposed model are twofold. First, a latent variational module (LVM) is attached to a conformer-based encoder-decoder ASR backbon…
Conversational context information, higher-level knowledge that spans across sentences, can help to recognize a long conversation. However, existing speech recognition models are typically built at a sentence level, and thus it may not capture important conversational context information. The recent progress in end-to-end speech recognition enables integrating context with other available information (e.g., acoustic, linguistic resources) and directly recognizing words from speech. In this work, we present a direct acoustic-to-word, end-to-end speech recognition model capable of utilizing the …
Carnegie Mellon University（卡内基梅隆大学）；Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院）
Spoken language translation applications for speech suffer due to conversational speech phenomena, particularly the presence of disfluencies. With the rise of end-to-end speech translation models, processing steps such as disfluency removal that were previously an intermediate step between speech recognition and machine translation need to be incorporated into model architectures. We use a sequence-to-sequence model to translate from noisy, disfluent speech to fluent text with disfluencies removed using the recently collected `copy-edited' references for the Fisher Spanish-English dataset. We …
