Neural Natural Language Generation: A Survey on Multilinguality, Multimodality, Controllability and Learning
Authors: Erkut Erdem, Menekse Kuyu, Semih Yagcioglu, Anette Frank, Letitia Parcalabescu, Barbara Plank, Andrii Babii, Oleksii Turuta, Aykut Erdem, Iacer Calixto, Elena Lloret, Elena-Simona Apostol, Ciprian-Octavian Truică, Branislava Šandrih, Sanda Martinčić-Ipšić, Gábor Berend, Albert Gatt, Grăzina Korvel
JAIR 2022 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | In this state-of-the-art report, we investigate the recent developments and applications of NNLG in its full extent from a multidimensional view, covering critical perspectives such as multimodality, multilinguality, controllability and learning strategies. We summarize the fundamental building blocks of NNLG approaches from these aspects and provide detailed reviews of commonly used preprocessing steps and basic neural architectures. This report also focuses on the seminal applications of these NNLG models such as machine translation, description generation, automatic speech recognition, abstractive summarization, text simplification, question answering and generation, and dialogue generation. Finally, we conclude with a thorough discussion of the described frameworks by pointing out some open research directions. |
| Researcher Affiliation | Academia | Erkut Erdem EMAIL Menekse Kuyu EMAIL Semih Yagcioglu EMAIL Hacettepe University, Ankara, Turkey Anette Frank EMAIL Letitia Parcalabescu EMAIL Heidelberg University, Heidelberg, Germany Andrii Babii EMAIL Oleksii Turuta EMAIL Kharkiv National University of Radio Electronics, Ukraine Aykut Erdem EMAIL Ko c University, Istanbul, Turkey Iacer Calixto EMAIL New York University, U.S.A. University of Amsterdam, Netherlands Barbara Plank EMAIL LMU Munich, Germany IT University of Copenhagen, Copenhagen, Denmark Elena Lloret EMAIL University of Alicante, Alicante, Spain Elena-Simona Apostol EMAIL Ciprian-Octavian Truic a EMAIL University Politehnica of Bucharest, Bucharest, Romania Branislava ˇSandrih EMAIL University of Belgrade, Belgrade, Serbia Sanda Martinˇci c-Ipˇsi c EMAIL University of Rijeka, Rijeka, Croatia G abor Berend EMAIL University of Szeged, Szeged, Hungary Albert Gatt EMAIL Utrecht University, The Netherlands University of Malta, Malta Graˇzina Korvel EMAIL Vilnius University, Vilnius, Lithuania |
| Pseudocode | No | The paper describes various neural architectures using mathematical equations (e.g., Equation 1-9 in Section 5), but it does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide concrete access to source code for the methodology described in this paper. Footnote 2 states: "See https://github.com/Multi3Generation/neural-natural-language-generation for the list of official implementations of the papers (if any) reviewed in our survey." This refers to code for *other* papers, not the current survey paper's methodology. |
| Open Datasets | No | This is a survey paper that reviews existing research and does not conduct its own experiments using specific datasets. While it mentions many datasets used in other works (e.g., MS COCO, VQA v2.0, How2 dataset), it does not provide access information for a dataset used in *this* paper's methodology or experiments. |
| Dataset Splits | No | This is a survey paper that reviews existing research and does not conduct its own experiments. Therefore, it does not provide specific dataset split information for its own work. |
| Hardware Specification | No | This is a survey paper that reviews existing research and does not conduct its own experiments. Therefore, it does not specify any hardware details used for running experiments. |
| Software Dependencies | No | This is a survey paper that reviews existing research and does not conduct its own experiments. Therefore, it does not provide specific ancillary software details with version numbers. |
| Experiment Setup | No | This is a survey paper that reviews existing research and does not conduct its own experiments. Therefore, it does not provide specific experimental setup details, hyperparameters, or training configurations. |