Website of George Pashev (Jidai Mirai)
Scientist, Programmer, Data Scientist, Entrepreneur
Foreign Language Assessment with Speech Translation & Multimodal Analytics
Welcome to the future of language assessment! This presentation introduces an innovative framework for foreign language assessment that integrates advanced speech-to-speech translation and multimodal learning analytics. Developed by George Pashev and Silvia Gaftandzhieva from the University of Plovdiv "Paisii Hilendarski," this system leverages the latest advancements in Natural Language Processing (NLP) and cognitive science to provide a comprehensive and adaptive evaluation tool.
Key Features:
- Speech-to-Speech Translation: Real-time translation capabilities support multilingual assessment across over 60 languages.
- Multimodal Learning Analytics: Combines acoustic, linguistic, temporal, and cognitive metrics to provide a holistic view of learner performance.
- Bloom's Taxonomy Integration: Aligns assessments with educational objectives, ensuring that evaluations target different levels of cognitive complexity.
- Adaptive Learning: Adjusts task difficulty in real-time based on learner performance, promoting personalized learning pathways.
- Scalable and Inclusive: Designed for global accessibility, supporting diverse learner populations and ensuring educational equity.
Presentation Overview:
Introduction: The need for a unified model combining pedagogical soundness with real-time speech processing.
System Overview: A detailed look at the 7-tuple model integrating language support, cognitive levels, and multimodal analytics.
Architecture: Modular design principles ensuring scalability, maintainability, and extensibility.
Test Generation: Structured templates aligned with Bloom's cognitive levels for comprehensive assessment.
Evaluation: Feature normalization, weighted aggregation, and interpretable feedback for meaningful assessments.
Results & Impact: Current progress, expert agreement, and system validation metrics.
Conclusion: Theoretical contributions and future directions for empirical validation and cross-domain applications.
Why This Matters:
This research bridges the gap between cognitive science, linguistics, and educational technology, offering a theoretical and practical foundation for next-generation language learning systems.
Presentation Link Here: - (3531) Foreign Language Assessment with Speech Translation & Multimodal Analytics - YouTube