---
title: "Design and Comparison of Arabic Negotiation Bots Using LLMs"
description: "The paper builds what the authors present as the first Arabic negotiation bots. 14 references."
authors: ["Ahmad Hajj", "Yasmine A Abu Adla", "Samah Albast", "Hazem Hajj", "Shady Elbassuoni", "Wassim El Hajj", "Khaled Shaban"]
year: 2024
venue: "Proceedings of the 7th International Conference on Natural Language and Speech Processing (ICNLSP 2024)"
source: "https://aclanthology.org/2024.icnlsp-1.33/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/hajj-2024-design-comparison-arabic-negotiation-bots/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/hajj-2024-design-comparison-arabic-negotiation-bots/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2024](https://md.research.elamur.ai/free/en/papers/2024/page/16/) / Design and Comparison of Arabic Negotiation Bots Using LLMs versus Seq2Seq Models with Reinforcement Learning

# Design and Comparison of Arabic Negotiation Bots Using LLMs versus Seq2Seq Models with Reinforcement Learning

**Authors:** Ahmad Hajj, [Yasmine A Abu Adla](https://md.research.elamur.ai/free/en/authors/yasmine-a-abu-adla/), Samah Albast, [Hazem Hajj](https://md.research.elamur.ai/free/en/authors/hazem-hajj/), [Shady Elbassuoni](https://md.research.elamur.ai/free/en/authors/shady-elbassuoni/), [Wassim El Hajj](https://md.research.elamur.ai/free/en/authors/wassim-el-hajj/) and [Khaled Shaban](https://md.research.elamur.ai/free/en/authors/khaled-shaban/) · **Year:** 2024 · **Venue:** [Proceedings of the 7th International Conference on Natural Language and Speech Processing (ICNLSP 2024)](https://md.research.elamur.ai/free/en/venues/icnlsp/) ([International Conference on Natural Language and Speech Processing 2024](https://md.research.elamur.ai/free/en/venues/icnlsp/2024/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2024.icnlsp-1.33/) · [PDF](https://aclanthology.org/2024.icnlsp-1.33.pdf)

## Summary
The paper builds what the authors present as the first Arabic negotiation bots. English negotiation agents are well studied, but Arabic has had no comparable systems. The setup follows Meta's "Deal or No Deal" scenario, in which a buyer and a seller divide a shared pool of items (books, hats, balls) that each side values differently.

The authors compare two approaches:
- **LLM approach:** GPT-3.5 Turbo, driven by zero-shot Arabic prompts, plays both roles in self-play. It can also be asked to generate tactical advice for each side.
- **RNN-RL baseline:** a Seq2Seq GRU-RNN refined with reinforcement learning, trained on an Arabic version of Meta's dialogue corpus produced by machine translation.

Five human judges rated both systems. The LLM scored far higher on coherence, complexity, language quality, style and negotiation dynamics. However, it sometimes miscounted items and "split" objects that cannot be divided. The work matters because it supplies released prompts, an Arabic dataset and a reusable recipe for porting negotiation agents to lower-resource languages.

## Key points
- **Data:** Meta's English corpus of 5808 human-human dialogues from 2236 scenarios was translated with the Google Translate API. A native speaker checked that the meaning stayed understandable. The data was split 80/10/10 into train, validation and test.
- **LLM prompting:** GPT-3.5 Turbo ran zero-shot, with no fine-tuning. The prompts set up two named personas (seller Sameer, buyer Amani) and a fixed price range of $10–$20. Each turn was conditioned on the full dialogue history.
- **Advice feature:** the LLM reviewed earlier exchanges and produced four recommendations for the buyer or the seller. Negotiation dynamics were compared before and after this feedback.
- **Human evaluation (5 evaluators):** the LLM averaged 4.2–4.6 across all criteria (coherence 4.6). The RNN-RL baseline scored about 1–1.2.
- **Failures and future work:** observed problems included miscounted allocations and hallucinated splits of indivisible items. The authors also note that advice in Arabic was more generic than in English, that behaviour differed by role, and that the training data was imbalanced between buyers and sellers.

## Interesting ideas
- A single model plays both negotiating sides, and each reply is conditioned on the whole history, so full dialogues come out without any human partner.
- The LLM split a ball between the two parties, showing that fluent text can break basic physical facts such as an object's indivisibility.
- The LLM's allocations sometimes failed to add up to the starting item count, which the authors had to correct by hand.
- The Arabic corpus was made by machine-translating an English benchmark and then having a native speaker judge only whether the meaning came through, not whether the wording was perfect.
- The paper contrasts verbose LLM turns, rich in justifications and reasoning, with terse RNN-RL turns that only propose quantities, showing that negotiation moves come in distinct kinds.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): Mova's self-play could stage two readers that question and answer each other about one text, producing dialogue-style training and test material without manual annotation.
- **I2** (world): adding a "countable/indivisible" property to entities in the ontology would let the world model flag holdings like "half a ball" as impossible rather than accept them.
- **I3** (answer): a conservation check that sums all holdings after each transfer and compares the total to the initial stock would catch such errors in maths word problems and explain where the count went wrong.
- **I4** (back): when testing translation through the meta-world, the evaluation could likewise score whether the meaning survives rather than penalising imperfect wording.
- **I5** (frames): speech-act tagging in prag could separate offer, counteroffer and justification frames, so a reader can trace which utterance actually changed the deal.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

📚 Shelves (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works

🧭 Topics (1): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works

## Cite

Cite as: Ahmad Hajj, Yasmine A Abu Adla, Samah Albast, Hazem Hajj, Shady Elbassuoni, Wassim El Hajj and Khaled Shaban. 2024. Design and Comparison of Arabic Negotiation Bots Using LLMs versus Seq2Seq Models with Reinforcement Learning. Proceedings of the 7th International Conference on Natural Language and Speech Processing (ICNLSP 2024).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{hajj2024design,
  title = {{Design and Comparison of Arabic Negotiation Bots Using LLMs versus Seq2Seq Models with Reinforcement Learning}},
  author = {Ahmad Hajj, Yasmine A Abu Adla, Samah Albast, Hazem Hajj, Shady Elbassuoni, Wassim {El Hajj} and Khaled Shaban},
  booktitle = {Proceedings of the 7th International Conference on Natural Language and Speech Processing (ICNLSP 2024)},
  year = {2024},
  url = {https://aclanthology.org/2024.icnlsp-1.33/}
}
```

- [← previous Creating an Aligned Corpus of Sound and Text: The Multimodal Corpus of Shakespeare and Milton](https://md.research.elamur.ai/free/en/papers/agirrezabal-2024-creating-aligned-corpus-sound-text/)
- [next → DiaSet: An Annotated Dataset of Arabic Conversations](https://md.research.elamur.ai/free/en/papers/israeli-2024-diaset-annotated-dataset-arabic-conversations/)
- [↑ back to papers of 2024](https://md.research.elamur.ai/free/en/papers/2024/page/16/)

## References

14 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Bala M Balachandran and Masoud Mohammadian. 2015. Development of a fuzzy-based multi-agent system for e-commerce settings. Procedia Computer Science, 60:593–602. [find](https://scholar.google.com/scholar?q=Bala+M+Balachandran+and+Masoud+Mohammadian%2E+2015%2E+Development+of+a+fuzzy%2Dbased+multi%2Dagent+system+for+e%2Dcommerce+settings%2E+Procedia+Computer+Science%2C+60%3A593%E2%80%93602%2E)
2. Founders. 2022. Simulation labs. Accessed 9 Sept. 2023. [find](https://scholar.google.com/scholar?q=Founders%2E+2022%2E+Simulation+labs%2E+Accessed+9+Sept%2E+2023%2E)
3. Yao Fu, Hao Peng, Tushar Khot, and Mirella Lapata. 2023. [Improving language model negotiation with self-play and in-context learning from ai feedback](https://md.research.elamur.ai/free/en/papers/fu-2023-improving-language-model-negotiation-self/).
4. Blair Halver. 2022. A.i.Negotiator. Accessed on 9 Sept. 2023. [find](https://scholar.google.com/scholar?q=Blair+Halver%2E+2022%2E+A%2Ei%2ENegotiator%2E+Accessed+on+9+Sept%2E+2023%2E)
5. He He, Derek Chen, Anusha Balakrishnan, and Percy Liang. 2018. [Decoupling strategy and generation in negotiation dialogues](https://md.research.elamur.ai/free/en/papers/he-2018-decoupling-strategy-generation-negotiation-dialogues/). In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 2333–2343.
6. Mohamed Jawahar Hussain. 2014. Parallel processing and intercommunication of multiple negotiation processes in e-commerce for optimal business. In 2014 Fourth International Conference on Communication Systems and Network Technologies, pages 583–586. IEEE. [find](https://scholar.google.com/scholar?q=Mohamed+Jawahar+Hussain%2E+2014%2E+Parallel+processing+and+intercommunication+of+multiple+negotiation+processes+in+e%2Dcommerce+for+optimal+business%2E+In+2014+Fourth+International+Conference+on+Communication)
7. Nicholas R Jennings, Peyman Faratin, Alessio R Lomuscio, Simon Parsons, Carles Sierra, and Michael Wooldridge. 2001. Automated negotiation: prospects, methods and challenges. International Journal of Group Decision and Negotiation, 10(2):199–215. [find](https://scholar.google.com/scholar?q=Nicholas+R+Jennings%2C+Peyman+Faratin%2C+Alessio+R+Lomuscio%2C+Simon+Parsons%2C+Carles+Sierra%2C+and+Michael+Wooldridge%2E+2001%2E+Automated+negotiation%3A+prospects%2C+methods+and+challenges%2E+International+Journal+of+)
8. Gaurav Koley and Shrisha Rao. 2018. Adaptive human-agent multi-issue bilateral negotiation using the thomas-kilmann conflict mode instrument. In 2018 IEEE/ACM 22nd International Symposium on Distributed Simulation and Real Time Applications (DS-RT), pages 1–5. IEEE. [find](https://scholar.google.com/scholar?q=Gaurav+Koley+and+Shrisha+Rao%2E+2018%2E+Adaptive+human%2Dagent+multi%2Dissue+bilateral+negotiation+using+the+thomas%2Dkilmann+conflict+mode+instrument%2E+In+2018+IEEE%2FACM+22nd+International+Symposium+on+Distribut)
9. Mike Lewis, Denis Yarats, Yann Dauphin, Devi Parikh, and Dhruv Batra. 2017. [Deal or no deal? end-to-end learning of negotiation dialogues](https://md.research.elamur.ai/free/en/papers/lewis-2017-deal-or-no-deal-end/). In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2443–2453, Copenhagen, Denmark. Association for Computational Linguistics. [↗](https://doi.org/10.18653/v1/d17-1259)
10. Kshitij Mishra, Azlaan Mustafa Samad, Palak Totala, and Asif Ekbal. 2022. [Pepds: A polite and empathetic persuasive dialogue system for charity donation](https://md.research.elamur.ai/free/en/papers/mishra-2022-pepds-polite-empathetic-persuasive-dialogue/). In Proceedings of the 29th International Conference on Computational Linguistics, pages 424–440.
11. Aritra Raut, Abhisek Tiwari, Subrata Das, Sriparna Saha, Anutosh Maitra, Roshni Ramnani, and Shubhashis Sengupta. 2023. Reinforcing personalized persuasion in task-oriented virtual sales assistant. Plos one, 18(1):e0275750. [find](https://scholar.google.com/scholar?q=Aritra+Raut%2C+Abhisek+Tiwari%2C+Subrata+Das%2C+Sriparna+Saha%2C+Anutosh+Maitra%2C+Roshni+Ramnani%2C+and+Shubhashis+Sengupta%2E+2023%2E+Reinforcing+personalized+persuasion+in+task%2Doriented+virtual+sales+assistant%2E+Pl)
12. Xuewei Wang, Weiyan Shi, Richard Kim, Yoojung Oh, Sijia Yang, Jingwen Zhang, and Zhou Yu. 2019. [Persuasion for good: Towards a personalized persuasive dialogue system for social good](https://md.research.elamur.ai/free/en/papers/wang-2019-persuasion-good-personalized-persuasive-dialogue/). In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5635–5649, Florence, Italy. Association for Computational Linguistics. [↗](https://doi.org/10.18653/v1/p19-1566)
13. Ronald J Williams. 1992. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine learning, 8:229–256. [find](https://scholar.google.com/scholar?q=Ronald+J+Williams%2E+1992%2E+Simple+statistical+gradient%2Dfollowing+algorithms+for+connectionist+reinforcement+learning%2E+Machine+learning%2C+8%3A229%E2%80%93256%2E)
14. Yiheng Zhou, He He, Alan W Black, and Yulia Tsvetkov. 2019. A dynamic strategy coach for effective negotiation. In Proceedings of the 20th Annual SIGdial Meeting on Discourse and Dialogue, pages 367–378, Stockholm, Sweden. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Yiheng+Zhou%2C+He+He%2C+Alan+W+Black%2C+and+Yulia+Tsvetkov%2E+2019%2E+A+dynamic+strategy+coach+for+effective+negotiation%2E+In+Proceedings+of+the+20th+Annual+SIGdial+Meeting+on+Discourse+and+Dialogue%2C+pages+367%E2%80%933)

*Updated 2 November 2026*