Bert onhttps://h30306.github.io/blog/tags/bert/Recent content in Bert onHugo -- gohugo.ioenhwork0511@gmail.com (Wei Yi Chung)hwork0511@gmail.com (Wei Yi Chung)© 2026 Wei Yi ChungFri, 01 Nov 2024 00:00:00 +0000Dynamic Traffic Network Representation Model for MRT Passenger Flowhttps://h30306.github.io/blog/paper/2024-kbs-dynamic-traffic-network-representation/Fri, 01 Nov 2024 00:00:00 +0000hwork0511@gmail.com (Wei Yi Chung)https://h30306.github.io/blog/paper/2024-kbs-dynamic-traffic-network-representation/Knowledge-Based Systems 2024: BERT dynamic station representations plus a hierarchical attention model that predicts MRT passenger flow from time, station and route viewsA Chinese Dimensional Valence-Arousal-Irony Detection on Sentence-level and Context-level Using Deep Learning Modelhttps://h30306.github.io/blog/paper/2022-ijclclp-cdvai-irony-detection/Fri, 30 Dec 2022 00:00:00 +0000hwork0511@gmail.com (Wei Yi Chung)https://h30306.github.io/blog/paper/2022-ijclclp-cdvai-irony-detection/IJCLCLP 2022: the journal extension of CDVAI, applying three feature-combination setups at both the sentence and context level and adding automatic prediction of the ironic context spanFactor Detection Task of Cyberbullying Using the Deep Learning Modelhttps://h30306.github.io/blog/paper/2022-ieee-bigdata-cyberbullying-factor-detection/Sat, 17 Dec 2022 00:00:00 +0000hwork0511@gmail.com (Wei Yi Chung)https://h30306.github.io/blog/paper/2022-ieee-bigdata-cyberbullying-factor-detection/IEEE BigData 2022 Workshop: defining six cyberbullying factors, building a Chinese PTT corpus, and using BERT to detect them and analyze bullying patterns around public figuresA Dimensional Valence-Arousal-Irony Dataset for Chinese Sentence and Contexthttps://h30306.github.io/blog/paper/2022-rocling-cdvai-irony-dataset/Mon, 21 Nov 2022 00:00:00 +0000hwork0511@gmail.com (Wei Yi Chung)https://h30306.github.io/blog/paper/2022-rocling-cdvai-irony-dataset/ROCLING 2022: extending the NTU irony corpus with sentence-level V/A/I and context-level V/A intensities, and benchmarking it with BERTDVA-BERT: Sentiment-Based Masked Language Modeling for Valence–Arousal Predictionhttps://h30306.github.io/blog/paper/2022-applied-intelligence-dva-bert/Tue, 22 Mar 2022 00:00:00 +0000hwork0511@gmail.com (Wei Yi Chung)https://h30306.github.io/blog/paper/2022-applied-intelligence-dva-bert/Applied Intelligence 2022: masking sentiment words by their intensity instead of random tokens, as an auxiliary task that improves sentence-level valence–arousal predictionA Dynamic Embedding Method for Passenger Flow Estimationhttps://h30306.github.io/blog/paper/2021-iiai-aai-dynamic-embedding-passenger-flow/Thu, 01 Jul 2021 00:00:00 +0000hwork0511@gmail.com (Wei Yi Chung)https://h30306.github.io/blog/paper/2021-iiai-aai-dynamic-embedding-passenger-flow/IIAI-AAI 2021: BERT-pretrained dynamic MRT station embeddings replace GMAN’s static station features; Competitive Paper Award