检索结果-内蒙古大学图书馆

Measurement of Music Aesthetics Using Deep Neural Networks and Dissonances

INFORMATION 2023年第7期14卷 358页

作者： Paroiu, Razvan Trausan-Matu, Stefan Univ Politehn Bucuresti Comp Sci & Engn Dept 313 Splaiul Independentei Bucharest 060042 Romania Romanian Acad Res Inst Artificial Intelligence Bucharest 050711 Romania Acad Romanian Scientists Str Ilfov 3 Bucharest 050044 Romania

In this paper, a new method that computes the aesthetics of a melody fragment is proposed, starting from dissonances. While music generated with artificial intelligence applications may be produced considerably more quickly than human-composed music, it has the drawback of not being appreciated like a human composition, being many times perceived by humans as artificial. For achieving supervised machine learning objectives of improving the quality of the great number of generated melodies, it is a challenge to ask humans to grade them. Therefore, it would be preferable if the aesthetics of artificial-intelligence-generated music is calculated by an algorithm. The proposed method in this paper is based on a neural network and a mathematical formula, which has been developed with the help of a study in which 108 students evaluated the aesthetics of several melodies. For evaluation, numerical values generated by this method were compared with ratings provided by human listeners from a second study in which 30 students participated and scores were generated by an existing different method developed by psychologists and three other methods developed by musicians. Our method achieved a Pearson correlation of 0.49 with human aesthetic scores, which is a much better result than other methods obtained. Additionally, our method made a distinction between human-composed melodies and artificial-intelligence-generated scores in the same way that human listeners did.

关键词： aesthetics computational aesthetics music generation dissonances deep neural networks sequence-to-sequence

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Approach for Automatic Romanian Lemmatization 25

Deep Learning Approach for Automatic Romanian Lemmatization

引用

25th KES International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES)

作者： Nutu, Maria Babes Bolyai Univ Fac Math & Comp Sci Dept Comp Sci RO-400084 Cluj Napoca Romania Tech Univ Cluj Napoca Commun Dept RO-400114 Cluj Napoca Romania

This paper proposes a deep learning sequence-to-sequence approach to improve the task of automatic Romanian lemmatization. The study compares 24 systems using different combinations of recurrent, convolutional and attention layers, while the text input consists of word-lemma pairs, both at word and trigram level. As Romanian is a low resourced language in the field of text processing, the aim of this study is to use as little input information as possible. Thus, to increase the lemmatization accuracy, additional lexical features (such as context and POS tags) have been provided only gradually. For the trigrams case, two scenarios were proposed: to predict the lemma for every word from the sequence or to predict the lemma only for the word in the middle. The lemmatizers have been analyzed on two Romanian datasets: the Romanian Explicative Dictionary (DEX) and the belletristic subset of the CoRoLa corpus. For the DEX dataset, the best results were obtained with the LSTM-based systems at both word (99.32%) and character level (99.43%). For the CoRoLa subset, the CNN-based architecture outperforms at trigram (95.86%) and word level (99.09%) while the LSTM-stacked system obtained the highest accuracy at character level (98.78%). (C) 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://***/licenses/by-nc-nd/4.0) Peer-review under responsibility of the scientific committee of KES International.

关键词： Lemmatization Part-of-Speech tagging LSTM CNN deep neural networks sequence-to-sequence Romanian

来源：评论

学校读者我要写书评

暂无评论

Stylized Dialogue Generation

Stylized Dialogue Generation

引用

IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)

作者： Ke, Shih-Wen Chen, Wei-Liang Natl Cent Univ Dept Informat Management Taoyuan Taiwan

ISBN: (纸本)9781665437714

Dialogue systems such as intelligent online customer services, online chatbots or smart kiosks are becoming increasingly popular. Currently dialogue systems lack personality and ability to respond according to contexts. In this study, we propose an approach to transfer the text into multiple styles when generating dialogue responses. It is especially challenging to build a stylized dialogue system as it combines supervised and unsupervised tasks. In practice, the dialogue data are usually paired, i.e. query paired response while styled text is not. Therefore, we propose using lightweight deep neural network models to bridge the dialogue response generation model and the style transfer model. This structure allows the model to generate responses of different styles to the same query. Our approach will be evaluated against selected state-of-the-art dialogue generation and style transfer techniques.

关键词： Dialogue generation style transfer sequence-to-sequence deep learning

来源：评论

学校读者我要写书评

暂无评论

RNN-T MODELS FAIL TO GENERALIZE TO OUT-OF-DOMAIN AUDIO: CAUSES AND SOLUTIONS

RNN-T MODELS FAIL TO GENERALIZE TO OUT-OF-DOMAIN AUDIO: CAUS...

引用

IEEE Spoken Language Technology Workshop (SLT)

作者： Chiu, Chung-Cheng Narayanan, Arun Han, Wei Prabhavalkar, Rohit Zhang, Yu Jaitly, Navdeep Pang, Ruoming Sainath, Tara N. Nguyen, Patrick Cao, Liangliang Wu, Yonghui Google Inc DE Shaw Grp Mountain View CA 94043 USA

ISBN: (纸本)9781728170664

In recent years, all-neural end-to-end approaches have obtained state-of-the-art results on several challenging automatic speech recognition (ASR) tasks. However, most existing works focus on building ASR models where train and test data are drawn from the same domain. This results in poor generalization characteristics on mismatched-domains: e.g., end-to-end models trained on short segments perform poorly when evaluated on longer utterances. In this work, we analyze the generalization properties of streaming and non-streaming recurrent neural network transducer (RNN-T) based end-to-end models in order to identify model components that negatively affect generalization performance. We propose two solutions: combining multiple regularization techniques during training, and using dynamic overlapping inference. On a long-form YouTube test set, when the non-streaming RNN-T model is trained with shorter segments of data, the proposed combination improves word error rate (WER) from 22:3% to 14:8%;when the streaming RNN-T model trained on short Search queries, the proposed techniques improve WER on the YouTube set from 67:0% to 25:3%. Finally, when trained on Librispeech, we find that dynamic overlapping inference improves WER on YouTube from 99:8% to 33:0%

关键词： Speech recognition RNN-T end-to-end sequence-to-sequence long-form

来源：评论

学校读者我要写书评

暂无评论

Reducing Streaming ASR Model Delay with Self Alignment 22

Reducing Streaming ASR Model Delay with Self Alignment

引用

Interspeech Conference

作者： Kim, Jaeyoung Lu, Han Tripathi, Anshuman Zhang, Qian Sak, Hasim Google Inc Mountain View CA 94043 USA

ISBN: (纸本)9781713836902

Reducing prediction delay for streaming end-to-end ASR models with minimal performance regression is a challenging problem. Constrained alignment is a well-known existing approach that penalizes predicted word boundaries using external low-latency acoustic models. On the contrary, recently proposed FastEmit is a sequence-level delay regularization scheme encouraging vocabulary tokens over blanks without any reference alignments. Although all these schemes are successful in reducing delay, ASR word error rate (WER) often severely degrades after applying these delay constraining schemes. In this paper, we propose a novel delay constraining method, named self alignment. Self alignment does not require external alignment models. Instead, it utilizes Viterbi forced-alignments from the trained model to find the lower latency alignment direction. From LibriSpeech evaluation, self alignment outperformed existing schemes: 25% and 56% less delay compared to FastEmit and constrained alignment at the similar word error rate. For Voice Search evaluation, 12% and 25% delay reductions were achieved compared to FastEmit and constrained alignment with more than 2% WER improvements.

关键词： Transformer RNN-T sequence-to-sequence encoder-decoder end-to-end speech recognition

来源：评论

学校读者我要写书评

暂无评论

FANS: Fusing ASR and NLU for on-device SLU 22

FANS: Fusing ASR and NLU for on-device SLU

引用

Interspeech Conference

作者： Radfar, Martin Mouchtaris, Athanasios Kunzmann, Siegfried Rastrow, Ariya Amazon Alexa Machine Learning Seattle WA 98109 USA

ISBN: (纸本)9781713836902

Spoken language understanding (SLU) systems translate voice input commands to semantics which are encoded as an intent and pairs of slot tags and values. Most current SLU systems deploy a cascade of two neural models where the first one maps the input audio to a transcript (ASR) and the second predicts the intent and slots from the transcript (NLU). In this paper, we introduce FANS, a new end-to-end SLU model that fuses an ASR audio encoder to a multi-task NLU decoder to infer the intent, slot tags, and slot values directly from a given input audio, obviating the need for transcription. FANS consists of a shared audio encoder and three decoders, two of which are seq-to-seq decoders that predict non null slot tags and slot values in parallel and in an auto-regressive manner. FANS neural encoder and decoders architectures are flexible which allows us to leverage different combinations of LSTM, self-attention, and attenders. Our experiments show compared to the state-of-the-art end-to-end SLU models, FANS reduces ICER and IRER errors relatively by 30% and 7%, respectively, when tested on an in-house SLU dataset and by 0.86% and 2% absolute when tested on a public SLU dataset.

关键词： Spoken language understanding (SLU) Automatic speech recognition (ASR) Natural language understanding (NLU) sequence-to-sequence neural transformer encoder-decoder intent-slot

来源：评论

学校读者我要写书评

暂无评论

EAT: ENHANCED ASR-TTS FOR SELF-SUPERVISED SPEECH RECOGNITION

EAT: ENHANCED ASR-TTS FOR SELF-SUPERVISED SPEECH RECOGNITION

引用

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

作者： Baskar, Murali Karthick Burget, Lukas Watanabe, Shinji Astudillo, Ramon Fernandez Cernocky, Jan Honza Brno Univ Technol Brno Czech Republic Johns Hopkins Univ Baltimore MD USA IBM Res Yorktown Hts NY USA

ISBN: (纸本)9781728176055

Self-supervised ASR-TTS models suffer in out-of-domain data conditions. Here we propose an enhanced ASR-TTS (EAT) model that incorporates two main features: 1) The ASR -> TTS direction is equipped with a language model reward to penalize the ASR hypotheses before forwarding it to TTS. 2) In the TTS -> ASR direction, a hyper-parameter is introduced to scale the attention context from synthesized speech before sending it to ASR to handle out-of-domain data. Training strategies and the effectiveness of the EAT model are explored under out-of-domain data conditions. The results show that EAT reduces the performance gap between supervised and self-supervised training significantly by absolute 2.6% and 2.7% on Librispeech and BABEL respectively.

关键词： cycle-consistency self-supervision sequence-to-sequence speech recognition

来源：评论

学校读者我要写书评

暂无评论

SHOW AND SPEAK: DIRECTLY SYNTHESIZE SPOKEN DESCRIPTION OF IMAGES

SHOW AND SPEAK: DIRECTLY SYNTHESIZE SPOKEN DESCRIPTION OF IM...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

作者： Wang, Xinsheng Feng, Siyuan Zhu, Jihua Hasegawa-Johnson, Mark Scharenborg, Odette Xi An Jiao Tong Univ Sch Software Engn Xian Peoples R China Delft Univ Technol Multimedia Comp Grp Delft Netherlands Univ Illinois Dept Elect & Comp Engn Urbana IL USA

ISBN: (纸本)9781728176055

This paper proposes a new model, referred to as the show and speak (SAS) model that, for the first time, is able to directly synthesize spoken descriptions of images, bypassing the need for any text or phonemes. The basic structure of SAS is an encoder-decoder architecture that takes an image as input and predicts the spectrogram of speech that describes this image. The final speech audio is obtained from the predicted spectrogram via WaveNet. Extensive experiments on the public benchmark database Flickr8k demonstrate that the proposed SAS is able to synthesize natural spoken descriptions for images, indicating that synthesizing spoken descriptions for images while bypassing text and phonemes is feasible.

关键词： Image-to-speech image captioning speech synthesis sequence-to-sequence encoder-decoder

来源：评论

学校读者我要写书评

暂无评论

Balancing Speed and Accuracy in Neural-Enhanced Phonetic Name Matching

Balancing Speed and Accuracy in Neural-Enhanced Phonetic Nam...

引用

21st Joint European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD)

作者： Blair, Philip Eliav, Carmel Hasanaj, Fiona Bar, Kfir Basis Technol 1060 Broadway Somerville NJ USA

ISBN: (纸本)9783030865177;9783030865160

Automatic co-text free name matching has a variety of important real-world applications, ranging from fiscal compliance to border control. Name matching systems use a variety of engines to compare two names for similarity, with one of the most critical being phonetic name similarity. In this work, we re-frame existing work on neural sequence-to-sequence transliteration such that it can be applied to name matching. Subsequently, for performance reasons, we then build upon this work to utilize an alternative, non-recurrent neural encoder module. This ultimately yields a model which is 63% faster while still maintaining a 16% improvement in averaged precision over our baseline model.

关键词： Name matching Transliteration Natural language processing sequence-to-sequence Multilingual Performance

来源：评论

学校读者我要写书评

暂无评论

Improving Text Summarization Using Feature Extraction Approach Based on Pointer-generator with Coverage 1

引用

18th Web Information Systems and Applications Conference (WISA)

作者： Chen, Yongchao He, Xin Wang, Guanghui Yu, Junyang Henan Prov Engn Res Ctr Intelligent Data Proc Kaifeng Peoples R China Henan Univ Henan Int Joint Lab Intelligent Network Theory & Kaifeng Peoples R China

ISBN: (数字)9783030875718

ISBN: (纸本)9783030875718;9783030875701

The sequence-to-sequence models have been widely used in the abstractive summarization tasks. However, the existing models have inaccuracy issues due to the insufficient understanding of overall semantics, the emergence of Out-Of-Vocabulary (OOV) words and the duplication of generated summarizations. In order to solve the above issues, in this paper, we propose an automatic text summarization model through combining the feature extraction, pointer-generator and coverage mechanism. First, in order to obtain input text with the multiple semantics, a feature extraction method is proposed to extract hidden features in the source text. Then, the pointer-generator (PGN) is used to select words from the source text or vocabulary to form a summarization, which solves the problem of OOV words in the generated summarizations. Finally, the coverage mechanism is applied to reduce the duplication of generated summarizations. The experimental results show that the model achieves better performance on the real-world CNN/Daily Mail summarization dataset.

关键词： Abstractive summarization sequence-to-sequence Feature extraction Coverage mechanism

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：