Challenges and Opportunities in Text Generation Explainability

被引:0
|
作者
Amara, Kenza [1 ]
Sevastjanow, Rita [1 ]
El-Assady, Mennatallah [1 ]
机构
[1] Swiss Fed Inst Technol, Dept Comp Sci, Zurich, Switzerland
关键词
Explainability; Text generation; Autoregressive models; Evaluation; Perturbation-based analysis; Challenges and opportunities;
D O I
10.1007/978-3-031-63787-2_13
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The necessity for interpretability in natural language processing (NLP) has risen alongside the growing prominence of large language models. Among the myriad tasks within NLP, text generation stands out as a primary objective of autoregressive models. The NLP community has begun to take a keen interest in gaining a deeper understanding of text generation, leading to the development of model-agnostic explainable artificial intelligence (xAI) methods tailored to this task. The design and evaluation of explainability methods are non-trivial since they depend on many factors involved in the text generation process, e.g., the autoregressive model and its stochastic nature. This paper outlines 17 challenges categorized into three groups that arise during the development and assessment of attribution-based explainability methods. These challenges encompass issues concerning tokenization, defining explanation similarity, determining token importance and prediction change metrics, the level of human intervention required, and the creation of suitable test datasets. The paper illustrates how these challenges can be intertwined, showcasing new opportunities for the community. These include developing probabilistic word-level explainability methods and engaging humans in the explainability pipeline, from the data design to the final evaluation, to draw robust conclusions on xAI methods.
引用
收藏
页码:244 / 264
页数:21
相关论文
共 50 条
  • [1] SYNTAXSHAP: Syntax-aware Explainability Method for Text Generation
    Amara, Kenza
    Sevastjanova, Rita
    El-Assady, Mennatallah
    FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: ACL 2024, 2024, : 4551 - 4566
  • [2] Pediatrics in Artificial Intelligence Era: A Systematic Review on Challenges, Opportunities, and Explainability
    Balla, Yashaswini
    Tirunagari, Santosh
    Windridge, David
    INDIAN PEDIATRICS, 2023, 60 (07) : 561 - 569
  • [3] Pediatrics in Artificial Intelligence Era: A Systematic Review on Challenges, Opportunities, and Explainability
    Yashaswini Balla
    Santosh Tirunagari
    David Windridge
    Indian Pediatrics, 2023, 60 : 561 - 569
  • [4] Erratum to: Pediatrics in Artificial Intelligence Era: A Systematic Review on Challenges, Opportunities, and Explainability
    Yashaswini Balla
    Santosh Tirunagari
    David Windridge
    Indian Pediatrics, 2024, 61 (1) : 65 - 65
  • [5] In-Time Explainability in Multi-Agent Systems: Challenges, Opportunities, and Roadmap
    Alzetta, Francesco
    Giorgini, Paolo
    Najjar, Amro
    Schumacher, Michael, I
    Calvaresi, Davide
    EXPLAINABLE, TRANSPARENT AUTONOMOUS AGENTS AND MULTI-AGENT SYSTEMS (EXTRAAMAS 2020), 2020, 12175 : 39 - 53
  • [6] Opportunities and challenges of text mining in materials research
    Kononova, Olga
    He, Tanjin
    Huo, Haoyan
    Trewartha, Amalie
    Olivetti, Elsa A.
    Ceder, Gerbrand
    ISCIENCE, 2021, 24 (03)
  • [7] Opportunities and Challenges in Next Generation Standards
    Stage, E. K.
    Asturias, H.
    Cheuk, T.
    Daro, P. A.
    Hampton, S. B.
    SCIENCE, 2013, 340 (6130) : 276 - 277
  • [8] Explainability Engineering Challenges: Connecting Explainability Levels to Run-Time Explainability
    Schwammberger, Maike
    Mirandola, Raffaela
    Wenninghoff, Nils
    EXPLAINABLE ARTIFICIAL INTELLIGENCE, XAI 2024, PT IV, 2024, 2156 : 205 - 218
  • [9] Opportunities and Challenges of Text Input in Portable Virtual Reality
    Knierim, Pascal
    Kosch, Thomas
    Groschopp, Johannes
    Schmidt, Albrecht
    CHI'20: EXTENDED ABSTRACTS OF THE 2020 CHI CONFERENCE ON HUMAN FACTORS IN COMPUTING SYSTEMS, 2020,
  • [10] A Systematic Review on Text Summarization: Techniques, Challenges, Opportunities
    Sharma, Kanta Prasad
    Yajid, Mohd Shukri Ab
    Gowrishankar, J.
    Mahajan, Rohini
    Alsoud, Anas Ratib
    Jadhav, Abhilasha
    Singh, Devendra
    EXPERT SYSTEMS, 2025, 42 (04)