{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T10:04:55Z","timestamp":1763719495068,"version":"3.45.0"},"reference-count":40,"publisher":"MDPI AG","issue":"12","license":[{"start":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T00:00:00Z","timestamp":1763683200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62266033"],"award-info":[{"award-number":["62266033"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Open Project of the Key Laboratory of Infinite-Dimensional Hamiltonian Systems and Their Algorithmic Applications (Ministry of Education) at Inner Mongolia Normal University","award":["2023KFZD03"],"award-info":[{"award-number":["2023KFZD03"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Computers"],"abstract":"<jats:p>Multi-faceted controllable text generation can be viewed as an extension and combination of controllable text generation tasks. It requires the generation of fluent text while controlling multiple different attributes (e.g., negative emotions and environmental protection in themes). Current research either estimates compact latent spaces for multiple attributes, reducing interference between different attributes but making it difficult to control the balance between multiple attributes, or controls the balance between multiple attributes but requires complex searches for decoding. Based on these issues, we propose a new method called MacHa, which trains an attribute latent space using multiple loss functions and establishes a mapping between the attribute latent space and attributes in sentences using a VAE network. An energy model based on the Hamilton function is defined in the potential space to control the balance between multiple attributes. Subsequently, in order to reduce the complexity of the decoding process, we extract samples using the RL sampling method and send them to the VAE decoder to generate the final text. The experimental results show that the MacHa method generates text with higher accuracy than the baseline models after balancing multiple attributes and has a fast decoding speed.<\/jats:p>","DOI":"10.3390\/computers14120503","type":"journal-article","created":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T09:50:26Z","timestamp":1763718626000},"page":"503","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["MacHa: Multi-Aspect Controllable Text Generation Based on a Hamiltonian System"],"prefix":"10.3390","volume":"14","author":[{"given":"Delong","family":"Xu","sequence":"first","affiliation":[{"name":"College of Computer Science and Technology, Inner Mongolia Normal University, No. 81, Zhao Wuda Road, Saihan District, Hohhot 010022, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Min","family":"Lin","sequence":"additional","affiliation":[{"name":"College of Computer Science and Technology, Inner Mongolia Normal University, No. 81, Zhao Wuda Road, Saihan District, Hohhot 010022, China"},{"name":"Infinite-Dimensional a Hamiltonian System and Algorithms Application Key Laboratory of the Ministry of Education, Inner Mongolia Applied Mathematics Center, No. 81, Zhao Wuda Road, Saihan District, Hohhot 010022, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yurong","family":"Wang","sequence":"additional","affiliation":[{"name":"College of Computer Science and Technology, Inner Mongolia Normal University, No. 81, Zhao Wuda Road, Saihan District, Hohhot 010022, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2025,11,21]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Pei, J., Yang, K., and Klein, D. (2023). PREADD: Prefix-adaptive decoding for controlled text generation. arXiv.","DOI":"10.18653\/v1\/2023.findings-acl.636"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"250","DOI":"10.5715\/jnlp.31.250","article-title":"Focused prefix tuning for controllable text generation","volume":"31","author":"Ma","year":"2024","journal-title":"J. Nat. Lang. Process."},{"key":"ref_3","first-page":"14542","article-title":"Controlled text generation as continuous optimization with multiple constraints","volume":"34","author":"Kumar","year":"2021","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Huang, X., Liu, Z., Li, P., Li, T., Sun, M., and Liu, Y. (2022). An extensible plug-and-play method for multi-aspect controllable text generation. arXiv.","DOI":"10.18653\/v1\/2023.acl-long.849"},{"key":"ref_5","unstructured":"Becker, J., Wahle, J.P., Gipp, B., and Ruas, T. (2024). Text generation: A systematic literature review of tasks, evaluation, and challenges. arXiv."},{"key":"ref_6","unstructured":"Xie, Z. (2017). Neural text generation: A practical guide. arXiv."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"208","DOI":"10.1016\/j.aiopen.2023.08.012","article-title":"GPT understands, too","volume":"5","author":"Liu","year":"2024","journal-title":"AI Open"},{"key":"ref_8","unstructured":"Zhu, C., Liu, Y., Lyu, C., Yang, X., Chen, G., Wang, L., Luo, W., and Zhang, K. (2025). Towards Lightweight, Adaptive and Attribute-Aware Multi-Aspect Controllable Text Generation with Large Language Models. arXiv."},{"key":"ref_9","unstructured":"Konen, K., Jentzsch, S., Diallo, D., Sch\u00fctt, P., Bensch, O., Baff, R.E., Opitz, D., and Hecking, T. (2024). Style vectors for steering generative large language model. arXiv."},{"key":"ref_10","unstructured":"Liu, S., Ye, H., Xing, L., and Zou, J. (2023). In-context vectors: Making in context learning more effective and controllable through latent space steering. arXiv."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Krause, B., Gotmare, A.D., McCann, B., Keskar, N.S., Joty, S., Socher, R., and Rajani, N.F. (2020). Gedi: Generative discriminator guided sequence generation. arXiv.","DOI":"10.18653\/v1\/2021.findings-emnlp.424"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Yang, K., and Klein, D. (2021). FUDGE: Controlled text generation with future discriminators. arXiv.","DOI":"10.18653\/v1\/2021.naacl-main.276"},{"key":"ref_13","unstructured":"Sitdikov, A., Balagansky, N., Gavrilov, D., and Markov, A. (2022). Classifiers are better experts for controllable text generation. arXiv."},{"key":"ref_14","unstructured":"Mudgal, S., Lee, J., Ganapathy, H., Li, Y., Wang, T., Huang, Y., Chen, Z., Cheng, H.T., Collins, M., and Strohman, T. (2023). Controlled decoding from language models. arXiv."},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Zhong, Q., Ding, L., Liu, J., Du, B., and Tao, D. (2024). ROSE doesn\u2019t do that: Boosting the safety of instruction-tuned large language models with reverse prompt contrastive decoding. arXiv.","DOI":"10.18653\/v1\/2024.findings-acl.814"},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Choi, D., Kim, J., Gim, M., Lee, J., and Kang, J. (2024, January 21\u201325). DeepClair: Utilizing Market Forecasts for Effective Portfolio Selection. Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, Boise, ID, USA.","DOI":"10.1145\/3627673.3680008"},{"key":"ref_17","unstructured":"Maas, A., Daly, R.E., Pham, P.T., Huang, D., Ng, A.Y., and Potts, C. (2011, January 19\u201324). Learning word vectors for sentiment analysis. Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, Portland, OR, USA."},{"key":"ref_18","first-page":"649","article-title":"Character-level convolutional networks for text classification","volume":"28","author":"Zhang","year":"2015","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_19","unstructured":"Ajwani, R.D., Zhu, Z., Rose, J., and Rudzicz, F. (2024). Plug and Play with Prompts: A Prompt Tuning Approach for Controlling Text Generation. arXiv."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Li, W., Wei, W., Xu, K., Xie, W., Chen, D., and Cheng, Y. (2024). Reinforcement learning with token-level feedback for controllable text generation. arXiv.","DOI":"10.18653\/v1\/2024.findings-naacl.111"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"He, X. (2021). Parallel refinements for lexically constrained text generation with bart. arXiv.","DOI":"10.18653\/v1\/2021.emnlp-main.681"},{"key":"ref_22","unstructured":"Zhang, H., Wang, X., Li, C., Ao, X., and He, Q. (March, January 25). Controlling large language models through concept activation vectors. Proceedings of the AAAI Conference on Artificial Intelligence, Philadelphia, PA, USA."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Liu, Y., Liu, X., Zhu, X., and Hu, W. (2024). Multi-aspect controllable text generation with disentangled counterfactual augmentation. arXiv.","DOI":"10.18653\/v1\/2024.acl-long.500"},{"key":"ref_24","unstructured":"Huang, Y., Chen, D., and Umrawal, A.K. (2025). JAM: Controllable and Responsible Text Generation via Causal Reasoning and Latent Vector Manipulation. arXiv."},{"key":"ref_25","unstructured":"Chen, Z., Feng, M., Yan, J., and Zha, H. (2022). Learning neural Hamiltonian dynamics: A methodological overview. arXiv."},{"key":"ref_26","unstructured":"Zaremba, W., Sutskever, I., and Vinyals, O. (2014). Recurrent neural network regularization. arXiv."},{"key":"ref_27","unstructured":"Lee, D.D., Pham, P., Largman, Y., and Ng, A. (2009). Advances in neural information processing systems 22. Tech. Rep."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"L012031","DOI":"10.1103\/PhysRevResearch.6.L012031","article-title":"Learning hamiltonian neural koopman operator and simultaneously sustaining and discovering conservation laws","volume":"6","author":"Zhang","year":"2024","journal-title":"Phys. Rev. Res."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Ding, H., Pang, L., Wei, Z., Shen, H., Cheng, X., and Chua, T.S. (2023). Maclasa: Multi-aspect controllable text generation via efficient sampling from compact latent space. arXiv.","DOI":"10.18653\/v1\/2023.findings-emnlp.292"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Gu, Y., Feng, X., Ma, S., Zhang, L., Gong, H., and Qin, B. (2022). A distributional lens for multi-aspect controllable text generation. arXiv.","DOI":"10.18653\/v1\/2022.emnlp-main.67"},{"key":"ref_31","first-page":"30416","article-title":"Differentiable sampling of categorical distributions using the catlog-derivative trick","volume":"36","author":"Sansone","year":"2023","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"229","DOI":"10.1023\/A:1022672621406","article-title":"Simple statistical gradient-following algorithms for connectionist reinforcement learning","volume":"8","author":"Williams","year":"1992","journal-title":"Mach. Learn."},{"key":"ref_33","unstructured":"Dathathri, S., Madotto, A., Lan, J., Hung, J., Frank, E., Molino, P., Yosinski, J., and Liu, R. (2019). Plug and play language models: A simple approach to controlled text generation. arXiv."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Liu, A., Sap, M., Lu, X., Swayamdipta, S., Bhagavatula, C., Smith, N.A., and Choi, Y. (2021). DExperts: Decoding-time controlled text generation with experts and anti-experts. arXiv.","DOI":"10.18653\/v1\/2021.acl-long.522"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Hallinan, S., Liu, A., Choi, Y., and Sap, M. (2022). Detoxifying text with marco: Controllable revision with experts and anti-experts. arXiv.","DOI":"10.18653\/v1\/2023.acl-short.21"},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Mireshghallah, F., Goyal, K., and Berg-Kirkpatrick, T. (2022). Mix and match: Learning-free controllable text generation using energy language models. arXiv.","DOI":"10.18653\/v1\/2022.acl-long.31"},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"Qian, J., Dong, L., Shen, Y., Wei, F., and Chen, W. (2022). Controllable natural language generation with contrastive prefixes. arXiv.","DOI":"10.18653\/v1\/2022.findings-acl.229"},{"key":"ref_38","doi-asserted-by":"crossref","unstructured":"Liu, G., Feng, Z., Gao, Y., Yang, Z., Liang, X., Bao, J., He, X., Cui, S., Li, Z., and Hu, Z. (2022). Composable text controls in latent space with ODEs. arXiv.","DOI":"10.18653\/v1\/2023.emnlp-main.1030"},{"key":"ref_39","doi-asserted-by":"crossref","unstructured":"Liu, Z., Lin, W., Shi, Y., and Zhao, J. (2021, January 13\u201315). A robustly optimized BERT pre-training approach with post-training. Proceedings of the China National Conference on Chinese Computational Linguistics, Hohhot, China.","DOI":"10.1007\/978-3-030-84186-7_31"},{"key":"ref_40","first-page":"2579","article-title":"Visualizing data using t-SNE","volume":"9","author":"Maaten","year":"2008","journal-title":"J. Mach. Learn. Res."}],"container-title":["Computers"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2073-431X\/14\/12\/503\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,21]],"date-time":"2025-11-21T10:00:30Z","timestamp":1763719230000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2073-431X\/14\/12\/503"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,21]]},"references-count":40,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2025,12]]}},"alternative-id":["computers14120503"],"URL":"https:\/\/doi.org\/10.3390\/computers14120503","relation":{},"ISSN":["2073-431X"],"issn-type":[{"value":"2073-431X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,21]]}}}