{"type":"ScholarlyArticle","title":"CounterGeDi: A Controllable Approach to Generate Polite, Detoxified and Emotional Counterspeech","authors":[{"name":"Punyajoy Saha","slug":"punyajoy-saha","given":"Punyajoy","family":"Saha","affiliation":"Indian Institute of Technology , Kharagpur , India"},{"name":"Kanishk Singh","slug":"kanishk-singh","given":"Kanishk","family":"Singh","affiliation":"Indian Institute of Technology , Kharagpur , India"},{"name":"Adarsh Kumar","slug":"adarsh-kumar","given":"Adarsh","family":"Kumar","affiliation":"Indian Institute of Technology , Kharagpur , India"},{"name":"Binny Mathew","slug":"binny-mathew","given":"Binny","family":"Mathew","affiliation":"Indian Institute of Technology , Kharagpur , India"},{"name":"Animesh Mukherjee","slug":"animesh-mukherjee","given":"Animesh","family":"Mukherjee","affiliation":"Indian Institute of Technology , Kharagpur , India"}],"abstract":"Recently, many studies have tried to create generation models to assist counter speakers by providing counterspeech suggestions for combating the explosive proliferation of online hate. However, since these suggestions are from a vanilla generation model, they might not include the appropriate properties required to counter a particular hate speech instance. In this paper, we propose CounterGeDi - an ensemble of generative discriminators (GeDi) to guide the generation of a DialoGPT model toward more polite, detoxified, and emotionally laden counterspeech. We generate counterspeech using three datasets and observe significant improvement across different attribute scores. The politeness and detoxification scores increased by around 15% and 6% respectively, while the emotion in the counterspeech increased by at least 10% across all the datasets. We also experiment with triple-attribute control and observe significant improvement over single attribute results when combining complementing attributes, e.g., politeness, joyfulness and detoxification. In all these experiments, the relevancy of the generated text does not deteriorate due to the application of these controls.","doi":"10.24963/ijcai.2022/716","url":"https://ijcai.org/proceedings/2022/716","pdf":"https://ijcai.org/proceedings/2022/0716.pdf","markdown":"https://ijcai.org/proceedings/2022/716.md","bibtex":"https://ijcai.org/proceedings/2022/bibtex/716","volume":{"slug":"2022","title":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2022","url":"https://ijcai.org/proceedings/2022/","year":2022,"location":"Vienna","dates":"23-29 July 2022","editors":"Luc De Raedt","isbn":"978-1-956792-00-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2022"},"section":"Research Papers","track":"AI for Good","number":716,"pages":{"first":5157,"last":5163},"keywords":["Multidisciplinary Topics and Applications: Social Sciences","Natural Language Processing: Dialogue and Interactive Systems","Natural Language Processing: Language Generation"],"videos":[{"label":"Video #1","url":"https://slideslive.com/38987133","length_sec":78,"provider":"slideslive"},{"label":"Video #2","url":"https://slideslive.com/38985276","length_sec":646,"provider":"slideslive"}],"references":[{"position":1,"raw":"[Bartlett and Krasodomski-Jones, 2015] Jamie Bartlett and Alex Krasodomski-Jones. Counter-speech examining content that challenges extremism online. DEMOS, October, 2015.","title":"Counter-speech examining content that challenges extremism online","authors":"Krasodomski-Jones ; Bartlett, Jamie Bartlett, Alex Krasodomski, -Jones","year":"2015","venue":null,"doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Bender et al., 2021] Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. On the dangers of stochastic parrots: Can language models be too big? In FAccT, 2021.","title":"On the dangers of stochastic parrots: Can language models be too big?","authors":"Emily M Bender, Timnit Gebru, Angelina Mcmillan-Major, Shmargaret Shmitchell","year":"2021","venue":"FAccT","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Benesch et al., 2016] Susan Benesch, Derek Ruths, Kelly P Dil- lon, Haji Mohammad Saleem, and Lucas Wright. Considerations for successful counterspeech. A report for Public Safety Canada under the Kanishka Project. Accessed November, 25:2020, 2016.","title":"Considerations for successful counterspeech. A report for Public Safety Canada under the Kanishka Project","authors":"Susan Benesch, Derek Ruths, Haji Kelly P Dillon, Lucas Mohammad Saleem, Wright","year":"2016","venue":null,"doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Chung et al., 2019] Yi-Ling Chung, Elizaveta Kuzmenko, Serra Sinem Tekiroglu, and Marco Guerini. Conan-counter narratives through nichesourcing: a multilingual dataset of responses to fight online hate speech. In ACL, 2019.","title":"Conan-counter narratives through nichesourcing: a multilingual dataset of responses to fight online hate speech","authors":"Yi-Ling Chung, Elizaveta Kuzmenko, Serra Sinem Tekiroglu, Marco Guerini","year":"2019","venue":"ACL","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Clark et al., 2019] Leigh Clark, Nadia Pantidi, Orla Cooney, Philip Doyle, Diego Garaialde, Justin Edwards, Brendan Spillane, Emer Gilmartin, Christine Murad, Cosmin Munteanu, Vincent Wade, and Benjamin R. Cowan. What Makes a Good Conversation? Challenges in Designing Truly Conversational Agents. ACM, 2019.","title":"What Makes a Good Conversation? Challenges in Designing Truly Conversational Agents","authors":"Leigh Clark, Nadia Pantidi, Orla Cooney, Philip Doyle, Diego Garaialde, Justin Edwards, Brendan Spillane, Emer Gilmartin, Christine Murad, Cosmin Munteanu, Vincent Wade, Benjamin R Cowan","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Dathathri et al., 2019] Sumanth Dathathri, Andrea Madotto, Jan- ice Lan, Jane Hung, Eric Frank, Piero Molino, Jason Yosinski, and Rosanne Liu. Plug and play language models: A simple ap- proach to controlled text generation. In ICLP, 2019.","title":"Plug and play language models: A simple approach to controlled text generation","authors":"Sumanth Dathathri, Andrea Madotto, Janice Lan, Jane Hung, Eric Frank, Piero Molino, Jason Yosinski, Rosanne Liu","year":"2019","venue":"ICLP","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[de los Riscos and D'Haro, 2021] Agustín Manuel de los Riscos and Luis Fernando D'Haro. ToxicBot: A Conversational Agent to Fight Online Hate Speech. 2021.","title":"Agustín Manuel de los Riscos and Luis Fernando D'Haro. ToxicBot: A Conversational Agent to Fight Online Hate Speech","authors":"Riscos De Los, D' Haro","year":"2021","venue":null,"doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Fanton et al., 2021] Margherita Fanton, Helena Bonaldi, Serra Sinem Tekiroglu, and Marco Guerini. Human-in-the- loop for data collection: a multi-target counter narrative dataset to fight online hate speech. In NAACL, 2021.","title":"Human-in-theloop for data collection: a multi-target counter narrative dataset to fight online hate speech","authors":"Margherita Fanton, Helena Bonaldi, Serra Sinem Tekiroglu, Marco Guerini","year":"2021","venue":"NAACL","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Frenett and Dow, 2015] Ross Frenett and Moli Dow. One to one online interventions: A pilot cve methodology. Institute for Strategic Dialogue, 2015.","title":"Ross Frenett and Moli Dow. One to one online interventions: A pilot cve methodology","authors":"Dow Frenett","year":"2015","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Fumagalli, 2020] Corrado Fumagalli. Counterspeech and or- dinary citizens: How? when? Political Theory, page 0090591720984724, 2020.","title":"Counterspeech and ordinary citizens: How? when? Political Theory","authors":"Corrado Fumagalli","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Holtzman et al., 2020] Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, and Yejin Choi. The curious case of neural text degener- ation. In ICLP, 2020.","title":"The curious case of neural text degeneration","authors":"Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, Yejin Choi","year":"2020","venue":"ICLP","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Krause et al., 2020] Ben Krause, Akhilesh Deepak Gotmare, Bryan McCann, Nitish Shirish Keskar, Shafiq Joty, Richard Socher, and Nazneen Fatema Rajani. Gedi: Generative discriminator guided sequence generation. arXiv preprint arXiv:2009.06367, 2020.","title":"Gedi: Generative discriminator guided sequence generation","authors":"Ben Krause, Akhilesh Deepak Gotmare, Bryan Mccann, Nitish Shirish Keskar, Shafiq Joty, Richard Socher, Nazneen Fatema, Rajani","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Liu et al., 2021] Alisa Liu, Maarten Sap, Ximing Lu, Swabha Swayamdipta, Chandra Bhagavatula, Noah A. Smith, and Yejin Choi. DExperts: Decoding-time controlled text generation with experts and anti-experts. In ACL-IJCNLP, 2021.","title":"DExperts: Decoding-time controlled text generation with experts and anti-experts","authors":"Alisa Liu, Maarten Sap, Ximing Lu, Swabha Swayamdipta, Chandra Bhagavatula, Noah A Smith, Yejin Choi","year":"2021","venue":"ACL-IJCNLP","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Madaan et al., 2020] Aman Madaan, Amrith Setlur, Tanmay Parekh, Barnabás Póczos, Graham Neubig, Yiming Yang, Ruslan Salakhutdinov, Alan W Black, and Shrimai Prabhumoye. Polite- ness transfer: A tag and generate approach. In ACL, 2020.","title":"Politeness transfer: A tag and generate approach","authors":"Aman Madaan, Amrith Setlur, Tanmay Parekh, Barnabás Póczos, Graham Neubig, Yiming Yang, Ruslan Salakhutdinov, Alan W Black, Shrimai Prabhumoye","year":"2020","venue":"ACL","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Mathew et al., 2019] Binny Mathew, Punyajoy Saha, Hardik Tharad, Subham Rajgaria, Prajwal Singhania, Suman Kalyan Maity, Pawan Goyal, and Animesh Mukherjee. Thou shalt not hate: Countering online hate speech. In ICWSM, 2019.","title":"Thou shalt not hate: Countering online hate speech","authors":"Binny Mathew, Punyajoy Saha, Hardik Tharad, Subham Rajgaria, Prajwal Singhania, Kalyan Suman, Pawan Maity, Animesh Goyal, Mukherjee","year":"2019","venue":"ICWSM","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Mathew et al., 2021] Binny Mathew, Punyajoy Saha, Seid Muhie Yimam, Chris Biemann, Pawan Goyal, and Animesh Mukherjee. Hatexplain: A benchmark dataset for explainable hate speech de- tection. In AAAI, 2021.","title":"Hatexplain: A benchmark dataset for explainable hate speech detection","authors":"Binny Mathew, Punyajoy Saha, Seid Muhie Yimam, Chris Biemann, Pawan Goyal, Animesh Mukherjee","year":"2021","venue":"AAAI","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Prabhumoye et al., 2020] Shrimai Prabhumoye, Alan W Black, and Ruslan Salakhutdinov. Exploring controllable text genera- tion techniques. arXiv preprint arXiv:2005.01822, 2020.","title":"Shrimai Prabhumoye, Alan W Black, and Ruslan Salakhutdinov. Exploring controllable text generation techniques","authors":"Prabhumoye","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Prendinger and Ishizuka, 2005] Helmut Prendinger and Mitsuru Ishizuka. The empathic companion: A character-based interface that addresses users' affective states. App. AI, 2005.","title":"The empathic companion: A character-based interface that addresses users' affective states","authors":"Ishizuka Prendinger, Helmut Prendinger, Mitsuru Ishizuka","year":"2005","venue":"App. AI","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Qian et al., 2019] Jing Qian, Anna Bethke, Yinyin Liu, Elizabeth Belding, and William Yang Wang. A benchmark dataset for learning to intervene in online hate speech. In EMNLP-IJCNLP, 2019.","title":"A benchmark dataset for learning to intervene in online hate speech","authors":"Jing Qian, Anna Bethke, Yinyin Liu, Elizabeth Belding, William Yang, Wang","year":"2019","venue":"EMNLP-IJCNLP","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Sai et al., 2020] Ananya B Sai, Akash Kumar Mohankumar, and Mitesh M Khapra. A survey of evaluation metrics used for nlg systems. arXiv preprint arXiv:2008.12009, 2020.","title":"A survey of evaluation metrics used for nlg systems","authors":"Akash Ananya B Sai, Mitesh M Kumar Mohankumar, Khapra","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Saravia et al., 2018] Elvis Saravia, Hsien-Chi Toby Liu, Yen-Hao Huang, Junlin Wu, and Yi-Shin Chen. Carer: Contextualized affect representations for emotion recognition. In EMNLP, 2018.","title":"Carer: Contextualized affect representations for emotion recognition","authors":"Saravia","year":"2018","venue":"EMNLP","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Sennrich et al., 2016] Rico Sennrich, Barry Haddow, and Alexan- dra Birch. Controlling politeness in neural machine translation via side constraints. In NAACL 2016, 2016.","title":"Controlling politeness in neural machine translation via side constraints","authors":"Sennrich","year":"2016","venue":"NAACL 2016","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Tekiroglu et al., 2020] Serra Sinem Tekiroglu, Yi-Ling Chung, and Marco Guerini. Generating counter narratives against online hate speech: Data and strategies. In ACL, 2020.","title":"Generating counter narratives against online hate speech: Data and strategies","authors":"Serra Sinem Tekiroglu, Yi-Ling Chung, Marco Guerini","year":"2020","venue":"ACL","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Vidgen et al., 2019] Bertie Vidgen, Helen Margetts, and Alex Har- ris. How much online abuse is there. Alan Turing Institute, 2019.","title":"How much online abuse is there","authors":"Bertie Vidgen, Helen Margetts, Alex Harris","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Wang and Wan, 2018] Ke Wang and Xiaojun Wan. Sentigan: Gen- erating sentimental texts via mixture adversarial networks. In IJ- CAI, 2018.","title":"Sentigan: Generating sentimental texts via mixture adversarial networks","authors":"Wang, Wan ; ] Ke Wang, Xiaojun Wan","year":"2018","venue":"IJ-CAI","doi":null,"url":null,"internal_url":"/proceedings/2018/618"},{"position":26,"raw":"[Warstadt and Bowman, 2019] Alex Warstadt and Samuel R Bow- man. Linguistic analysis of pretrained sentence encoders with acceptability judgments. arXiv preprint arXiv:1901.03438, 2019.","title":"Linguistic analysis of pretrained sentence encoders with acceptability judgments","authors":"Alex Warstadt, Samuel R Bowman","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Zhang et al., 2020] Yizhe Zhang, Siqi Sun, Michel Galley, Yen- Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, and William B Dolan. Dialogpt: Large-scale generative pre- training for conversational response generation. In ACL: System Demonstrations, 2020.","title":"Dialogpt: Large-scale generative pretraining for conversational response generation","authors":"Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, William B Dolan","year":"2020","venue":"ACL: System Demonstrations","doi":null,"url":null,"internal_url":null},{"position":28,"raw":"[Zhu and Bhat, 2021] Wanzheng Zhu and Suma Bhat. Generate, prune, select: A pipeline for counterspeech generation against online hate speech. arXiv preprint arXiv:2106.01625, 2021.","title":"Generate, prune, select: A pipeline for counterspeech generation against online hate speech","authors":"Bhat ; ] Wanzheng Zhu, Suma Zhu, Bhat","year":"2021","venue":null,"doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}