Today, artificial language models are transforming how we read, write, communicate, and even think. In an era when we can converse with machines, making such technology possible depends on understanding human meaning-making and language processes. From the early days when talking machines existed only in imagination, to the rise of neural networks and today’s large-scale language models, the story of language modeling is also the story of our search for intelligence, meaning, and connection.
This three-week course program, spanning three weeks, is designed for participants from diverse disciplines and does not require prior technical knowledge. Across nine 50-minute classes, participants will learn comprehensively how language models emerged, how they work, and how they influence our daily lives and future.
Our instructor, Çiğdem Taş, graduated from Boğaziçi University, Department of English Language and Literature and completed her master’s degree in Cognitive Science at Yeditepe University. During her undergraduate education, she took courses in Latin, Arabic, Russian, and Greek alongside linguistics, psychology, and history, and during her graduate studies, she completed coursework in statistics, programming, machine learning, artificial intelligence, and data science. She currently works as a linguist at a global company and has over ten years of professional experience in translation, language, and communication. Proficient in Turkish, English, Russian, and Greek, Çiğdem has a deep interest in the intersections of language, cognition, and human experience. Her work explores the functions of language in identity, thought, and communication, as well as how the human brain constructs and interprets meaning, with a particular focus on language modeling through sensory perception and free word association.
The course will take place on October 7, 14, and 21, Tuesdays, from 7:30 PM to 10:30 PM (GMT+3) via Zoom. The total duration of the course program will be approximately 9 hours.
Week 1 (October 7)
- Introduction: Fundamentals of linguistics and language modeling
- Development Process: Modern linguistics, early attempts, symbolic and data-driven approaches
- Background: Systems, the distributional hypothesis, n-grams, and their limitations
- In the first week, we will explore how early language modeling efforts began, how they evolved through trial and error, and learn about the underlying mechanisms of these systems.
Week 2 (October 14)
- Foundations: Word associations, mental lexicons, semantic networks, and world knowledge
- Technologies: Word vectors and embeddings, bag-of-words models, transformers
- How Machines Learn: Artificial neural networks (deep learning, RNNs, applications), cultural biases in data
- The Transformer Breakthrough: Knowing enough, “attention is all you need,” from BERT to GPT
- In the second week, we will examine the foundational structures on which language models and related technologies are built, exploring how these systems have evolved through them. We will also discuss how machines are trained, what they learn, and what the broader implications of this process are.
Week 3 (October 21)
- Large and Small Language Models: Data, learning, and fine-tuning
- Challenges and Ethical Questions: What’s happening? Ethical issues (rights and regulations); bias, hallucination, misinformation, and the “death of creativity”
- What Comes Next? In-class discussion, new applications, philosophical reflections, and open questions
- In the third and final week, we will discuss what artificial intelligence and language models bring into—and take away from—our lives, exploring how they transform the way we think, create, and communicate. We will reflect on the challenges and ethical dilemmas they raise, engage in class discussions, and attempt to envision the future of human–machine interaction together.
The participation fee for the course program is 2050 TL per person (including VAT). Payment details will be shared via email with participants whose applications have been approved.
