Skip to content
Prof. (Dr.) Ritika Mehra
Prof. (Dr.)Ritika MehraPro-Vice Chancellor
Course Study Deck Platform Protected Viewer25 Slides

Deep Learning Architectures

Unit 4: Natural Language Processing & Recommendation Systems

Instructor: Prof. (Dr.) Ritika Mehra · Pro-Vice Chancellor & Professor of Computer Science

Academic Protection Guard Active

Read-only online viewer. File downloading, text copying, and screenshot capture are disabled to safeguard university courseware.

1Course Overview·
Unit 4: Part 1Course OverviewPreserved Date: 27 Oct 2026

Deep Learning Architectures

Unit 4: Natural Language Processing and Recommendation Systems

School of Engineering & Computing / Department of Computer Science
Comprehensive Lecture Series on Deep Learning & NLP Architectures
Text Preprocessing, Tokenization, Embedding Vectors & Recommendation Models
Architectural Diagram & FlowProtected Slide Graphic
Diagram for Deep Learning Architectures
Key Pedagogical Takeaway: Core master lecture covering modern natural language processing pipelines and recommender architectures.
1 / 25

Unit 4 Curriculum: Learning Objectives

Deep Learning Architectures · Master of Technology & Ph.D. Level

In deep learning applied to natural language processing, text cannot be processed directly as symbols. It requires a disciplined multi-stage transformation from unstructured characters into semantic dense vectors.

1Text Preprocessing Foundations
2Tokenization Paradigms (Word, Character, Subword, Sentence, N-gram)
3Subword Algorithms (BPE, WordPiece, SentencePiece)
4Transformer-Based Tokenization & Numerical Embedding IDs
5Stemming, Overstemming & Understemming
6Lemmatization & Dictionary Lemmas
7Stop-Word Removal & Sentiment Exceptions
8Industrial Applications & Deep Architecture Integration

Platform Viewing Policy

Academic Integrity & Distribution Terms

This presentation is protected academic study material for registered students of Prof. (Dr.) Ritika Mehra. Direct downloading, redistribution, print reproduction, and screen recording are restricted under academic copyright.

Interactive Features:

  • Full keyboard navigation (Arrows / Space)
  • Fullscreen presentation mode (Press F)
  • Topic-based outline jumping & search
Author: Prof. (Dr.) Ritika MehraVerified Archive