Shaoting Guo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Shaoting Guo has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Lei Li
Beijing University of Posts and Telecommunications, Beijing, China
4 shared publications - Yingying Luo
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
3 shared publications - Zhejun Zhang
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China
3 shared publications - Hengyi Shao
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing 100876, China
2 shared publications - Liang Dong
School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing 100876, China
2 shared publications - Lin Zhang
Beijing University of Posts and Telecommunications
2 shared publications
Latest Publications
- Open Access
- 2 January 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-19
Brain-model neural similarity reveals abstractive summarization performance
Deep language models (DLMs) have exhibited remarkable language understanding and generation capabilities, prompting researchers to explore the similarities between their internal mechanisms and human language cognitive processing. This study investigated the representational similarity (RS) between the abstractive summarization (ABS) models and the human brain and its correlation to the performance of ABS tasks. Specifically, representational similarity analysis (RSA) was used to measure the similarity between the representational patterns (RPs) of the BART, PEGASUS, and T5 models’ hidden layers and the human brain’s language RPs under different spatiotemporal conditions. Layer-wise ablation manipulation, including attention ablation and noise addition was employed to examine the hidden layers’ effect on model performance. The results demonstrate that as the depth of hidden layers increases, the models’ text encoding becomes increasingly similar to the human brain’s language RPs. Manipulating deeper layers leads to more substantial decline in summarization performance compared to shallower layers, highlighting the crucial role of deeper layers in integrating essential information. Notably, the study confirms the hypothesis that the hidden layers exhibiting higher similarity to human brain activity play a more critical role in model performance, with their correlations reaching statistical significance even after controlling for perplexity. These findings deepen our understanding of the cognitive mechanisms underlying language representations in DLMs and their neural correlates, potentially providing insights for optimizing and improving language models by aligning them with the human brain’s language-processing mechanisms.
- 2 January 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-19
Open AccessBrain-model neural similarity reveals abstractive summarization performance
Deep language models (DLMs) have exhibited remarkable language understanding and generation capabilities, prompting researchers to explore the similarities between their internal mechanisms and human language cognitive processing. This study investigated the representational similarity (RS) between the abstractive summarization (ABS) models and the human brain and its correlation to the performance of ABS tasks. Specifically, representational similarity analysis (RSA) was used to measure the similarity between the representational patterns (RPs) of the BART, PEGASUS, and T5 models’ hidden layers and the human brain’s language RPs under different spatiotemporal conditions. Layer-wise ablation manipulation, including attention ablation and noise addition was employed to examine the hidden layers’ effect on model performance. The results demonstrate that as the depth of hidden layers increases, the models’ text encoding becomes increasingly similar to the human brain’s language RPs. Manipulating deeper layers leads to more substantial decline in summarization performance compared to shallower layers, highlighting the crucial role of deeper layers in integrating essential information. Notably, the study confirms the hypothesis that the hidden layers exhibiting higher similarity to human brain activity play a more critical role in model performance, with their correlations reaching statistical significance even after controlling for perplexity. These findings deepen our understanding of the cognitive mechanisms underlying language representations in DLMs and their neural correlates, potentially providing insights for optimizing and improving language models by aligning them with the human brain’s language-processing mechanisms.
- Open Access
- 19 January 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (2) , 875
Exploring the Cognitive Neural Basis of Factuality in Abstractive Text Summarization Models: Interpretable Insights from EEG Signals
(1) Background: Information overload challenges decision-making in the Industry 4.0 era. While Natural Language Processing (NLP), especially Automatic Text Summarization (ATS), offers solutions, issues with factual accuracy persist. This research bridges cognitive neuroscience and NLP, aiming to improve model interpretability. (2) Methods: This research examined four fact extraction techniques: dependency relation, named entity recognition, part-of-speech tagging, and TF-IDF, in order to explore their correlation with human EEG signals. Representational Similarity Analysis (RSA) was applied to gauge the relationship between language models and brain activity. (3) Results: Named entity recognition showed the highest sensitivity to EEG signals, marking the most significant differentiation between factual and non-factual words with a score of −0.99. The dependency relation followed with −0.90, while part-of-speech tagging and TF-IDF resulted in 0.07 and −0.52, respectively. Deep language models such as GloVe, BERT, and GPT-2 exhibited noticeable influences on RSA scores, highlighting the nuanced interplay between brain activity and these models. (4) Conclusions: Our findings emphasize the crucial role of named entity recognition and dependency relations in fact extraction and demonstrate the independent effects of different models and TOIs on RSA scores. These insights aim to refine algorithms to reflect human text processing better, thereby enhancing ATS models’ factual integrity.
- 19 January 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (2) , 875
Open AccessExploring the Cognitive Neural Basis of Factuality in Abstractive Text Summarization Models: Interpretable Insights from EEG Signals
(1) Background: Information overload challenges decision-making in the Industry 4.0 era. While Natural Language Processing (NLP), especially Automatic Text Summarization (ATS), offers solutions, issues with factual accuracy persist. This research bridges cognitive neuroscience and NLP, aiming to improve model interpretability. (2) Methods: This research examined four fact extraction techniques: dependency relation, named entity recognition, part-of-speech tagging, and TF-IDF, in order to explore their correlation with human EEG signals. Representational Similarity Analysis (RSA) was applied to gauge the relationship between language models and brain activity. (3) Results: Named entity recognition showed the highest sensitivity to EEG signals, marking the most significant differentiation between factual and non-factual words with a score of −0.99. The dependency relation followed with −0.90, while part-of-speech tagging and TF-IDF resulted in 0.07 and −0.52, respectively. Deep language models such as GloVe, BERT, and GPT-2 exhibited noticeable influences on RSA scores, highlighting the nuanced interplay between brain activity and these models. (4) Conclusions: Our findings emphasize the crucial role of named entity recognition and dependency relations in fact extraction and demonstrate the independent effects of different models and TOIs on RSA scores. These insights aim to refine algorithms to reflect human text processing better, thereby enhancing ATS models’ factual integrity.
- Open Access
- 14 December 2023
- Published by Association for Computing Machinery (ACM)
- p. 107-111
Towards Robust Handwritten Tibetan Numeral Recognition with Spiking Neural Networks
The Tibetan region contains abundant cultural resources. Handwritten Tibetan numeral recognition is an important basic task for Tibetan language digitalization process. As a low-resource language, Tibetan poses more stringent robustness requirements for deep learning models, requiring them to resist interference and learn core features. This paper proposes a Spiking Fully Convolutional Neural network model SFCN based on spiking neural networks (SNNs) to accomplish handwritten Tibetan numeral recognition on the TibetanMNIST dataset. The fully convolutional neural network structure based on CNNs can effectively extract image features. Combined with SNNs as SFCN, it further enhances network robustness. Experiments show that our SFCN model maintains 98.90% accuracy while exhibiting extraordinary robustness under various interferences, compared to an artificial neural network (ANN) of the same structure. On the basis, combining interferences with training samples further improves model robustness. SFCN trained with adversarial attack method projected gradient descent (PGD) achieves the optimal robustness. SFCN-PGD model’s accuracy on the test set improved by 80.39% at the highest compared to traditional ANNs1.
- 14 December 2023
- Published by Association for Computing Machinery (ACM)
- p. 107-111
Open AccessTowards Robust Handwritten Tibetan Numeral Recognition with Spiking Neural Networks
The Tibetan region contains abundant cultural resources. Handwritten Tibetan numeral recognition is an important basic task for Tibetan language digitalization process. As a low-resource language, Tibetan poses more stringent robustness requirements for deep learning models, requiring them to resist interference and learn core features. This paper proposes a Spiking Fully Convolutional Neural network model SFCN based on spiking neural networks (SNNs) to accomplish handwritten Tibetan numeral recognition on the TibetanMNIST dataset. The fully convolutional neural network structure based on CNNs can effectively extract image features. Combined with SNNs as SFCN, it further enhances network robustness. Experiments show that our SFCN model maintains 98.90% accuracy while exhibiting extraordinary robustness under various interferences, compared to an artificial neural network (ANN) of the same structure. On the basis, combining interferences with training samples further improves model robustness. SFCN trained with adversarial attack method projected gradient descent (PGD) achieves the optimal robustness. SFCN-PGD model’s accuracy on the test set improved by 80.39% at the highest compared to traditional ANNs1.
- Open Access
- 27 November 2023
- Published by MDPI AG in Preprints.org
Exploring the Cognitive Neural Basis of Factuality in Abstractive Text Summarization Models: Interpretable Insights from EEG Signals
(1) Background: Information overload challenges decision-making in the Industry 4.0 era. While Natural Language Processing (NLP), especially Automatic Text Summarization (ATS), offers solutions, issues with factual accuracy persist. This research bridges cognitive neuroscience and NLP, aiming to improve model interpretability. (2) Methods: This research examined four fact extraction techniques: dependency relation, named entity recognition, part-of-speech tagging, and TF-IDF, in order to explore their correlation with human EEG signals. Representational Similarity Analysis (RSA) was applied to gauge the relationship between language models and brain activity. (3) Results: Named entity recognition showed the highest sensitivity to EEG signals, marking the most significant differentiation between factual and non-factual words with a score of -0.99. The dependency relation followed with -0.90, while part-of-speech tagging and TF-IDF resulted in 0.07 and -0.52, respectively. Deep language models such as GloVe, BERT, and GPT-2 exhibited noticeable influences on RSA scores, highlighting the nuanced interplay between brain activity and these models. (4) Conclusions: Our findings emphasize the crucial role of named entity recognition and dependency relations in fact extraction and demonstrate the independent effects of different models and TOIs on RSA scores. These insights aim to refine algorithms to reflect human text processing better, thereby enhancing ATS models’ factual integrity.
- 27 November 2023
- Published by MDPI AG in Preprints.org
Open AccessExploring the Cognitive Neural Basis of Factuality in Abstractive Text Summarization Models: Interpretable Insights from EEG Signals
(1) Background: Information overload challenges decision-making in the Industry 4.0 era. While Natural Language Processing (NLP), especially Automatic Text Summarization (ATS), offers solutions, issues with factual accuracy persist. This research bridges cognitive neuroscience and NLP, aiming to improve model interpretability. (2) Methods: This research examined four fact extraction techniques: dependency relation, named entity recognition, part-of-speech tagging, and TF-IDF, in order to explore their correlation with human EEG signals. Representational Similarity Analysis (RSA) was applied to gauge the relationship between language models and brain activity. (3) Results: Named entity recognition showed the highest sensitivity to EEG signals, marking the most significant differentiation between factual and non-factual words with a score of -0.99. The dependency relation followed with -0.90, while part-of-speech tagging and TF-IDF resulted in 0.07 and -0.52, respectively. Deep language models such as GloVe, BERT, and GPT-2 exhibited noticeable influences on RSA scores, highlighting the nuanced interplay between brain activity and these models. (4) Conclusions: Our findings emphasize the crucial role of named entity recognition and dependency relations in fact extraction and demonstrate the independent effects of different models and TOIs on RSA scores. These insights aim to refine algorithms to reflect human text processing better, thereby enhancing ATS models’ factual integrity.
- 29 October 2009
- Published by Elsevier in Radiation Measurements
- Vol. 44 (9-10) , 885-888
Study of bubble distributions by high-energy protons in bubble detectors and its hints in neutron detection at higher altitude and in space
- 29 October 2009
- Published by Elsevier in Radiation Measurements
- Vol. 44 (9-10) , 885-888
Study of bubble distributions by high-energy protons in bubble detectors and its hints in neutron detection at higher altitude and in space