Large language models (LLMs) have revolutionized the development of interactional artificial intelligence (AI) systems by democratizing their use. These models have shown remarkable advancements in various applications, such as conversational AI and machine translation, marking the undeniable advent of the human-machine era. However, despite their significant achievements, state-of-the-art systems still exhibit shortcomings in language understanding, raising questions about their true comprehension of human languages.

The concept of language understanding has always been contentious, as meaning-making depends not only on form and immediate meaning but also on context. Therefore, understanding natural language involves more than just parsing form and meaning; it requires access to grounding for true comprehension. Equipping language models with linguistics-grounded capabilities remains a complex task, given the importance of discourse, pragmatics, and social context in language understanding.

Assessing whether LLMs understand language is a doubly challenging task, as it necessitates not only grasping the intrinsic capabilities of LLMs but also examining their impact and requirements in real-world applications. While LLMs have shown effectiveness in various applications, the lack of supporting theories raises concerns about ethical implications, particularly in applications involving human interaction.

The “Language Understanding in the Human-Machine Era” (LUHME) workshop aims to reignite the debate on the role of understanding in natural language use and its applications. It seeks to explore the necessity of language understanding in computational tasks like machine translation and natural language generation, as well as the contributions of language professionals in enhancing computational language understanding.

Topics of Interest

Topics of interest include, but are not limited to:
- Language understanding in LLMs
- Language grounding
- Psycholinguistic approaches to language understanding
- Discourse, pragmatics and language understanding
- Intent detection
- Evaluation of language understanding
- Human vs. machine language understanding
- Machine translation/interpreting and language understanding
- Multimodality and language understanding
- Socio-cultural aspects in understanding language
- Effects and risks of language misunderstanding
- Manifestations of language (mis)understanding
- Natural language understanding and toxic content
- Ethical issues in language misunderstanding
- Distributional semantics and language understanding
- Linguistic theory and language understanding by machines
- Linguistic, world, and commonsense knowledge in language understanding
- Role of language professionals in the LLMs era
- Understanding language and explainable AI

For further information, please visit https://luhme.up.pt/