arxiv: 2605.11964 · v1 · submitted 2026-05-12 · 💻 cs.CL

Recognition: 2 theorem links

· Lean Theorem

Enhancing Target-Guided Proactive Dialogue Systems via Conversational Scenario Modeling and Intent-Keyword Bridging

Maodong Li , Yancui Li , Fang Kong

Authors on Pith no claims yet

Pith reviewed 2026-05-13 05:18 UTC · model grok-4.3

classification 💻 cs.CL

keywords target-guided dialogueproactive dialogue systemsconversational scenario modelingintent-keyword bridgingdialogue generationuser profilesdomain knowledge

0 comments

The pith

Jointly modeling user profiles and domain knowledge as conversational scenarios, combined with intent-keyword bridging, dynamically guides target-guided proactive dialogue systems toward pre-defined goals.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that existing target-guided dialogue systems fail to capture the shifting dynamics of real conversations because they overlook ongoing scenario modeling. By representing user profiles together with domain knowledge as a conversational scenario, the system generates a bias that continuously shapes its own utterances. It further adds intent-keyword bridging to forecast which keywords will be relevant in the next turns, supplying higher-level steering than single-turn keyword targets alone. Automatic and human evaluations are said to show clear gains in proactivity, fluency, and informativeness, closing the distance to natural interactions.

Core claim

Conversational scenario modeling treats user profiles and domain knowledge as a single source of bias that modulates every system utterance; intent-keyword bridging predicts the intent keywords that will be needed in upcoming turns and thereby supplies flexible, multi-turn guidance. When both components are active, the resulting dialogues are reported to be more proactive, fluent, and informative than those produced by prior target-guided systems.

What carries the argument

Conversational scenario modeling that fuses user profiles and domain knowledge into a dynamic bias, paired with intent-keyword bridging that forecasts future intent keywords.

If this is right

System utterances become continuously steered by an evolving scenario bias instead of reacting only to the immediate turn.
Predicted intent keywords give the system advance notice of what the user is likely to discuss next, enabling smoother topic transitions.
The combination raises proactivity, fluency, and informativeness scores while reducing the mismatch with real-world conversation patterns.
The approach narrows the performance gap between automated target-guided systems and human-led guided conversations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same scenario-plus-bridging structure could be reused in non-target-guided chatbots to improve long-term coherence without explicit goals.
If scenario modeling proves robust, it may reduce the need for hand-crafted dialogue policies in many goal-oriented systems.
Live deployment with actual users would test whether the predicted keywords remain accurate once the conversation leaves the training distribution.

Load-bearing premise

The new scenario bias and keyword bridging will accurately reflect how real conversations unfold and will reliably produce measurable gains over existing methods.

What would settle it

A controlled experiment in which the proposed model shows no statistically significant improvement over strong baselines on proactivity, fluency, or informativeness metrics when both are evaluated on the same target-guided dialogue tasks.

Figures

Figures reproduced from arXiv: 2605.11964 by Fang Kong, Maodong Li, Yancui Li.

**Figure 2.** Figure 2: Our framework. keyword-topic, and intent keyword, respectively. MP(·) denotes the max-pooling operation. By dynamically predicting m intent keywords and applying max pooling, we obtain the most relevant bridging intent keyword Hz for the next turn, considering the next m turns. The final step in our framework, generating the system utterance, can be represented as3 : rt = arg max P(rt |r<t, Hz , Hh , b)… view at source ↗

**Figure 3.** Figure 3: Variation of W. F1 w.r.t m and δ. On the DuRecDial2.0 dataset, fine-tuned LLMs show high target failure rates (e.g., LLaMA3B reaches 40.30% on the ID test set), indicating that fine-tuning alone is insufficient for mastering complex proactive planning. In contrast, large promptbased models (e.g., Qwen32B) achieve extremely low failure rates (3.80%) but often at the cost of dialogue quality, abruptly for… view at source ↗

**Figure 4.** Figure 4: Human evaluation. "App." denotes "appro [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

**Figure 5.** Figure 5: Case study (ID test set). (a) T5-Flan; (b) Our framework; (c) Intent keyword transitions. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗

**Figure 6.** Figure 6: Variation of Word F1 with respect to m and δ on the DuRecDial2.0 dataset. B Implementation Details We use AdamW as the optimizer and employ both a warmup strategy and gradient clipping. For our framework, training is performed for 50 epochs with a batch size of 8 and a learning rate of 3×10−5 . The hyperparameters m and the threshold δ are determined empirically; see §3.3 for details. We fine-tune LLaMA-1B… view at source ↗

**Figure 7.** Figure 7: Human evaluation on the DuRecDial dataset [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗

**Figure 8.** Figure 8: Case study (OOD test set). (a) Results of T5-Flan; (b) Results of our framework; (c) Dialogue-level intent [PITH_FULL_IMAGE:figures/full_fig_p018_8.png] view at source ↗

**Figure 9.** Figure 9: Our using prompt. 19 [PITH_FULL_IMAGE:figures/full_fig_p019_9.png] view at source ↗

read the original abstract

A target-guided proactive dialogue system aims to steer conversations proactively toward pre-defined targets, such as designated keywords or specific topics. During guided conversations, dynamically modeling conversational scenarios and intent keywords to guide system utterance generation is beneficial; however, existing work largely overlooks this aspect, resulting in a mismatch with the dynamics of real-world conversations. In this paper, we jointly model user profiles and domain knowledge as conversational scenarios to introduce a scenario bias that dynamically influences system utterances, and employ intent-keyword bridging to predict intent keywords for upcoming dialogue turns, providing higher level and more flexible guidance. Extensive automatic and human evaluations demonstrate the effectiveness of conversational scenario modeling and intent keyword bridging, yielding substantial improvements in proactivity, fluency, and informativeness for target-guided proactive dialogue systems, thereby narrowing the gap with real world interactions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper adds scenario modeling from profiles and domain knowledge plus intent-keyword bridging to make target-guided dialogues more adaptive, but the abstract gives no experimental details so the gains are hard to judge.

read the letter

The main contribution is a way to jointly encode user profiles and domain knowledge into conversational scenarios that bias the system's utterances on the fly, combined with an intent-keyword bridge that predicts keywords for upcoming turns. This targets the gap the authors identify between existing systems and the fluid nature of real conversations, aiming for more flexible, higher-level guidance than prior target-guided work. It is a straightforward extension rather than a big conceptual leap, but it does line up with practical needs in proactive dialogue. The approach is described clearly enough in the abstract to see how the two pieces fit together without obvious internal contradictions. The stress-test note is right that nothing jumps out as broken on the surface. The soft spot is the complete lack of methods, metrics, baselines, or data details. The abstract asserts substantial gains in proactivity, fluency, and informativeness from automatic and human evaluations, yet supplies nothing to check whether those gains are real or just fitting. That makes the soundness hard to assess from what is here. This paper is for NLP researchers working on goal-directed or proactive dialogue systems who want incremental ideas for handling conversation dynamics. A reader already familiar with the subfield could pick up the scenario-bias and bridging concepts and try them out, but would need the full experiments before treating the results as solid. I would send it to peer review because the problem is real and the framing is coherent; a referee can verify the implementation and numbers. If the experiments turn out thin, it can be revised rather than rejected outright.

Referee Report

2 major / 0 minor

Summary. The paper claims that jointly modeling user profiles and domain knowledge as conversational scenarios introduces a dynamic scenario bias to influence system utterances, while intent-keyword bridging predicts keywords for future turns to provide higher-level guidance. It reports that these components yield substantial gains in proactivity, fluency, and informativeness for target-guided proactive dialogue systems, as shown by extensive automatic and human evaluations that narrow the gap with real-world interactions.

Significance. If the empirical results hold under rigorous controls, the work could advance target-guided dialogue by supplying a more flexible, dynamic bias mechanism that better matches real conversation dynamics than prior static approaches. The dual use of automatic metrics and human judgments is a strength that would support broader adoption if baselines and datasets are clearly documented.

major comments (2)

[Abstract] Abstract: The abstract asserts 'substantial improvements' from 'extensive automatic and human evaluations' but supplies no methods, metrics, baselines, datasets, or implementation details. This absence makes it impossible to verify whether the evidence supports the central claims about effectiveness and real-world alignment.
[Introduction / Proposed Method] The weakest assumption—that scenario modeling and intent-keyword bridging will reliably capture real-world dynamics—remains untested in the provided description, as no controls for confounding factors or ablation studies are referenced.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment point by point below, offering clarifications from the full paper and proposing targeted revisions where the concerns are valid.

read point-by-point responses

Referee: [Abstract] Abstract: The abstract asserts 'substantial improvements' from 'extensive automatic and human evaluations' but supplies no methods, metrics, baselines, datasets, or implementation details. This absence makes it impossible to verify whether the evidence supports the central claims about effectiveness and real-world alignment.

Authors: Abstracts are conventionally concise and omit granular details such as specific metrics, baselines, and implementation to focus on high-level contributions and results. The full manuscript details the evaluation protocol in Sections 4 and 5, including automatic metrics for proactivity, fluency, and informativeness, the chosen baselines, the datasets employed, and implementation specifics. To address the concern, we will revise the abstract to include a brief clause referencing the evaluation framework and primary datasets used. revision: partial
Referee: [Introduction / Proposed Method] The weakest assumption—that scenario modeling and intent-keyword bridging will reliably capture real-world dynamics—remains untested in the provided description, as no controls for confounding factors or ablation studies are referenced.

Authors: The manuscript includes ablation studies (Section 4.3) that isolate the contributions of conversational scenario modeling and intent-keyword bridging, along with comparisons against multiple baselines that omit the dynamic bias mechanism to control for confounding factors. Human evaluations further assess alignment with real-world dynamics. We will add explicit forward references to these ablations and controls in the Introduction and Method sections to make their presence clearer. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected in modeling or evaluation chain

full rationale

The paper describes an architectural proposal: jointly encoding user profiles and domain knowledge as conversational scenarios that supply dynamic bias to utterance generation, plus an intent-keyword bridging mechanism to forecast keywords for future turns. These choices are presented as design decisions, followed by standard automatic and human evaluations that measure gains in proactivity, fluency, and informativeness. No equations, parameter-fitting steps, or derivation chains appear that reduce a claimed prediction or result to the inputs by construction. No load-bearing self-citations or uniqueness theorems imported from prior author work are invoked in the provided text. The central claims rest on empirical outcomes rather than definitional equivalence or renamed fits, rendering the approach self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The abstract contains no mathematical derivations, fitted parameters, axioms, or newly postulated entities; it describes high-level modeling techniques only.

pith-pipeline@v0.9.0 · 5436 in / 1020 out tokens · 49334 ms · 2026-05-13T05:18:31.865241+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

[1]

ACM Transactions on Information Systems , volume=

Target-constrained bidirectional planning for generation of target-oriented proactive dialogue , author=. ACM Transactions on Information Systems , volume=. 2024 , publisher=

work page 2024
[2]

IEEE Transactions on Neural Networks and Learning Systems , volume=

A target-driven planning approach for goal-directed dialog systems , author=. IEEE Transactions on Neural Networks and Learning Systems , volume=. 2023 , publisher=

work page 2023
[3]

2022 , eprint=

Scaling Instruction-Finetuned Language Models , author=. 2022 , eprint=

work page 2022
[4]

2024 , eprint=

The Llama 3 Herd of Models , author=. 2024 , eprint=

work page 2024
[5]

2025 , eprint=

Qwen2.5 Technical Report , author=. 2025 , eprint=

work page 2025
[6]

2024 , eprint=

DeepSeek LLM: Scaling Open-Source Language Models with Longtermism , author=. 2024 , eprint=

work page 2024
[7]

2023 , eprint=

Mistral 7B , author=. 2023 , eprint=

work page 2023
[8]

Enhancing Goal-oriented Proactive Dialogue Systems via Consistency Reflection and Correction

Zhang, Didi and Fan, Yaxin and Li, Peifeng and Zhu, Qiaoming. Enhancing Goal-oriented Proactive Dialogue Systems via Consistency Reflection and Correction. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025. doi:10.18653/v1/2025.acl-long.1050

work page doi:10.18653/v1/2025.acl-long.1050 2025
[9]

Interpersonal Memory Matters: A New Task for Proactive Dialogue Utilizing Conversational History

Wu, Bowen and Wang, Wenqing and Lihaoran, Lihaoran and Deng, Yunhan and Li, Ying and Yu, Jingsong and Wang, Baoxun. Interpersonal Memory Matters: A New Task for Proactive Dialogue Utilizing Conversational History. Proceedings of the 29th Conference on Computational Natural Language Learning. 2025. doi:10.18653/v1/2025.conll-1.4

work page doi:10.18653/v1/2025.conll-1.4 2025
[10]

and Liao, Lizi

Dao, Huy Quang and Deng, Yang and Bui, Khanh-Huyen and Le, Dung D. and Liao, Lizi. Experience as Source for Anticipation and Planning: Experiential Policy Learning for Target-driven Recommendation Dialogues. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024. doi:10.18653/v1/2024.findings-emnlp.829

work page doi:10.18653/v1/2024.findings-emnlp.829 2024
[11]

Enhancing Emotional Support Conversations: A Framework for Dynamic Knowledge Filtering and Persona Extraction

Hao, Jiawang and Kong, Fang. Enhancing Emotional Support Conversations: A Framework for Dynamic Knowledge Filtering and Persona Extraction. Proceedings of the 31st International Conference on Computational Linguistics. 2025

work page 2025
[12]

Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment

Xu, Kaishuai and Cheng, Yi and Hou, Wenjun and Tan, Qiaoyu and Li, Wenjie. Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment. Findings of the Association for Computational Linguistics: ACL 2024. 2024. doi:10.18653/v1/2024.findings-acl.406

work page doi:10.18653/v1/2024.findings-acl.406 2024
[13]

T op KG : Target-oriented Dialog via Global Planning on Knowledge Graph

Yang, Zhitong and Wang, Bo and Zhou, Jinfeng and Tan, Yue and Zhao, Dongming and Huang, Kun and He, Ruifang and Hou, Yuexian. T op KG : Target-oriented Dialog via Global Planning on Knowledge Graph. Proceedings of the 29th International Conference on Computational Linguistics. 2022

work page 2022
[14]

Target-Guided Open-Domain Conversation

Tang, Jianheng and Zhao, Tiancheng and Xiong, Chenyan and Liang, Xiaodan and Xing, Eric and Hu, Zhiting. Target-Guided Open-Domain Conversation. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. 2019. doi:10.18653/v1/P19-1565

work page doi:10.18653/v1/p19-1565 2019
[15]

Reinforced Target-driven Conversational Promotion

Dao, Huy and Liao, Lizi and Le, Dung and Nie, Yuxiang. Reinforced Target-driven Conversational Promotion. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.775

work page doi:10.18653/v1/2023.emnlp-main.775 2023
[16]

Dialogue Planning via Brownian Bridge Stochastic Process for Goal-directed Proactive Dialogue

Wang, Jian and Lin, Dongding and Li, Wenjie. Dialogue Planning via Brownian Bridge Stochastic Process for Goal-directed Proactive Dialogue. Findings of the Association for Computational Linguistics: ACL 2023. 2023. doi:10.18653/v1/2023.findings-acl.25

work page doi:10.18653/v1/2023.findings-acl.25 2023
[17]

Proceedings of the AAAI Conference on Artificial Intelligence , volume=

Keyword-guided neural conversational model , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

work page
[18]

Proceedings of the AAAI conference on artificial intelligence , volume=

Conceptnet 5.5: An open multilingual graph of general knowledge , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

work page
[19]

Advances in Neural Information Processing Systems , volume=

A simple language model for task-oriented dialogue , author=. Advances in Neural Information Processing Systems , volume=

work page
[20]

Proceedings of the AAAI Conference on Artificial Intelligence , volume=

A goal interaction graph planning framework for conversational recommendation , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

work page
[21]

IJCAI , year=

A Survey on Proactive Dialogue Systems: Problems, Methods, and Prospects , author=. IJCAI , year=

work page
[22]

ACM Transactions on Information Systems , volume=

Proactive conversational ai: A comprehensive survey of advancements and opportunities , author=. ACM Transactions on Information Systems , volume=. 2025 , publisher=

work page 2025
[23]

Towards Conversational Recommendation over Multi-Type Dialogs

Liu, Zeming and Wang, Haifeng and Niu, Zheng-Yu and Wu, Hua and Che, Wanxiang and Liu, Ting. Towards Conversational Recommendation over Multi-Type Dialogs. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020. doi:10.18653/v1/2020.acl-main.98

work page doi:10.18653/v1/2020.acl-main.98 2020
[24]

D u R ec D ial 2.0: A Bilingual Parallel Corpus for Conversational Recommendation

Liu, Zeming and Wang, Haifeng and Niu, Zheng-Yu and Wu, Hua and Che, Wanxiang. D u R ec D ial 2.0: A Bilingual Parallel Corpus for Conversational Recommendation. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. 2021. doi:10.18653/v1/2021.emnlp-main.356

work page doi:10.18653/v1/2021.emnlp-main.356 2021
[25]

KAPA : A Deliberative Agent Framework with Tree-Structured Knowledge Base for Multi-Domain User Intent Understanding

Tang, Jiakai and Shen, Shiqi and ZhipengWang, ZhipengWang and Zhi, Gong and Feng, Xueyang and Sun, Zexu and Tan, Haoran and Chen, Xu. KAPA : A Deliberative Agent Framework with Tree-Structured Knowledge Base for Multi-Domain User Intent Understanding. Findings of the Association for Computational Linguistics: ACL 2025. 2025. doi:10.18653/v1/2025.findings-acl.319

work page doi:10.18653/v1/2025.findings-acl.319 2025
[26]

The Moral Integrity Corpus: A Benchmark for Ethical Dialogue Systems

Ziems, Caleb and Yu, Jane and Wang, Yi-Chia and Halevy, Alon and Yang, Diyi. The Moral Integrity Corpus: A Benchmark for Ethical Dialogue Systems. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022. doi:10.18653/v1/2022.acl-long.261

work page doi:10.18653/v1/2022.acl-long.261 2022
[27]

Persuading across Diverse Domains: a Dataset and Persuasion Large Language Model

Jin, Chuhao and Ren, Kening and Kong, Lingzhen and Wang, Xiting and Song, Ruihua and Chen, Huan. Persuading across Diverse Domains: a Dataset and Persuasion Large Language Model. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024. doi:10.18653/v1/2024.acl-long.92

work page doi:10.18653/v1/2024.acl-long.92 2024
[28]

Proceedings of the 18th ACM Conference on Recommender Systems , pages=

Towards empathetic conversational recommender systems , author=. Proceedings of the 18th ACM Conference on Recommender Systems , pages=

work page
[29]

and Konstas, Ioannis and Rieser, Verena

Sevegnani, Karin and Howcroft, David M. and Konstas, Ioannis and Rieser, Verena. OTT ers: One-turn Topic Transitions for Open-Domain Dialogue. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021. doi:10.18653/v1/2021...

work page doi:10.18653/v1/2021.acl-long.194 2021
[30]

Advances in neural information processing systems , volume=

Attention is all you need , author=. Advances in neural information processing systems , volume=

work page
[31]

Multi-Hop Question Generation via Dual-Perspective Keyword Guidance

Li, Maodong and Zhang, Longyin and Kong, Fang. Multi-Hop Question Generation via Dual-Perspective Keyword Guidance. Findings of the Association for Computational Linguistics: ACL 2025. 2025. doi:10.18653/v1/2025.findings-acl.526

work page doi:10.18653/v1/2025.findings-acl.526 2025
[32]

, author=

Lora: Low-rank adaptation of large language models. , author=. ICLR , volume=

work page
[33]

A Diversity-Promoting Objective Function for Neural Conversation Models

Li, Jiwei and Galley, Michel and Brockett, Chris and Gao, Jianfeng and Dolan, Bill. A Diversity-Promoting Objective Function for Neural Conversation Models. Proceedings of the 2016 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies. 2016. doi:10.18653/v1/N16-1014

work page doi:10.18653/v1/n16-1014 2016
[34]

B leu: a method for automatic evaluation of machine translation

Papineni, Kishore and Roukos, Salim and Ward, Todd and Zhu, Wei-Jing. B leu: a Method for Automatic Evaluation of Machine Translation. Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics. 2002. doi:10.3115/1073083.1073135

work page doi:10.3115/1073083.1073135 2002
[35]

Journal of machine learning research , volume=

Exploring the limits of transfer learning with a unified text-to-text transformer , author=. Journal of machine learning research , volume=

work page
[36]

Journal of Machine Learning Research , volume=

Scaling instruction-finetuned language models , author=. Journal of Machine Learning Research , volume=

work page
[37]

T encent P retrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities

Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and Tian, Rong and Liu, Weijie and Chen, Yiren and Sun, Ningyuan and Liu, Haoyan and Mao, Weiquan and Guo, Han. T encent P retrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (...

work page doi:10.18653/v1/2023.acl-demo.20 2023
[38]

doi:10.57967/hf/2316 , publisher =

Wang, Shenzhi and Zheng, Yaowei and Wang, Guoyin and Song, Shiji and Huang, Gao , title =. doi:10.57967/hf/2316 , publisher =

work page doi:10.57967/hf/2316
[39]

doi:10.57967/hf/2317 , publisher =

Wang, Shenzhi and Zheng, Yaowei and Wang, Guoyin and Song, Shiji and Huang, Gao , title =. doi:10.57967/hf/2317 , publisher =

work page doi:10.57967/hf/2317
[40]

From Personas to Talks: Revisiting the Impact of Personas on LLM -Synthesized Emotional Support Conversations

Wu, Shenghan and Zhu, Yimo and Hsu, Wynne and Lee, Mong-Li and Deng, Yang. From Personas to Talks: Revisiting the Impact of Personas on LLM -Synthesized Emotional Support Conversations. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025. doi:10.18653/v1/2025.emnlp-main.277

work page doi:10.18653/v1/2025.emnlp-main.277 2025
[41]

ICASSP 2026-2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

Pseudo-Siamese Network for Planning in Target-Oriented Proactive Dialogues , author=. ICASSP 2026-2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2026 , organization=

work page 2026