Please ensure Javascript is enabled for purposes of website accessibility

Παρουσίαση/Προβολή

Εικόνα επιλογής

Special Topics in Language Technology II: Language and Cognitive Robotics

(M930) -  Katerina Pastra

Περιγραφή Μαθήματος

This course focuses on artificial cognitive systems and in particular on robots, with a special emphasis on the role of Natural Language within such systems. It is an interdisciplinary course which brings together theories, findings and methodologies from Cognitive Science, Cognitive Robotics, Neuroscience, Theoretical and Computational Linguistics and Multimedia Systems.

In the course, learners get acquainted with the idiosyncratic characteristics of natural language, perception and action and their semantic integration in multimodal everyday interaction and discourse. Step by step, they explore such integration through -among others- “syntax”, a fundamental cognitive mechanism that is core not only in verbal but also in sensorimotor behaviour and learn how syntax relates to long term memory. The latter involves getting familiar with a computational semantic memory module, in the form of a multimodal, referential, and recursive semantic network. The course is enriched with examples from two robotic applications: (a) verbal human-robot interaction in everyday life, and (b) visual scene understanding and verbalization. The main learning objective of the course is to introduce learners to an embodied perspective of Natural Language description and Engineering which renders reference a core requirement in language analysis and forms a bridge with sensorimotor behaviour.

Ημερομηνία δημιουργίας

Πέμπτη 10 Σεπτεμβρίου 2020

Σελίδα Μαθήματος
  • Περίγραμμα

    Course Objectives/Goals

    The course aims at providing students with basic knowledge on cognitive systems (cognitive robots in particular) and expose them to questions related to the role of language within such systems. In particular, the main learning objectives of the course comprise:  

    • Explaining what cognitive robots/systems are, the main challenges in developing such systems and the role language can play in such systems;
    • Describing the semantic characteristics and interrelations among language, perception and action and semantically analysing multisensory and multimodal communication;
    • Interpreting ‘syntax’ as a supra-modal, cognitive mechanism and applying a common syntactic framework for the analysis of both sensorymotor experiences and language;
    • Experiment with computational semantic memories in the form of multimodal, referential and recursive semantic networks.

    Upon successful completion of the course the students will be able to:

    1. Engage in discussion on the basic challenges in cognitive systems and be advocates of the role of language in them
    2. Semantically analyse multimodal/multisensory communication/material
    3. Use a common syntactic framework for the analysis of language and sensorimotor experience
    4. Use and enrich computational semantic memories

    Indicative Content

    • Introduction to Artificial Cognitive Systems and Cognitive Robotics
    • Cognitive Architectures and the role of Language
    • Language, Perception and Action (1): characteristics and relations
    • Language, Perception and Action (2): semantic interaction in multisensory and multimodal communication
    • Syntax as a Fundamental Cognitive Mechanism: from language grammar to the grammar of action
    • Recursion as a Cognitive Phenomenon
    • Embodied Language Processing
    • Computational Semantic Memory and the Role of Language
    • Robotic Applications: (a) Verbal Human-Robot Interaction: from language to action, and (b) Visual Scene Understanding: from image to language

    Assessment Methods

    • Course Participation: Students will be assessed for their active participation in class, comprising engagement in class discussion, and contribution to group activities during the ‘seminar’ part of the course.
    • Team Project I: Each team will choose a multimodal document/file for semantic annotation using an annotation tool and corresponding qualitative and quantitative analysis of findings. The work will be assessed through an oral presentation and a written report (approx. 2000 words).
    • Team Project II: Building on a “verbal human-robot interaction” scenario, student teams will proceed to using a robotic semantic memory for endowing the robot with the prior knowledge needed in interaction. The work will be assessed through an oral presentation and a written report (approx. 4000 words).

    Assessment methods

    Number

    Percentage

    Class participation

    1

    10%

    Team Project I

    2

    45%

    Team Project II

    3

    45%

    Textbooks - Reading List