My AI Journey

Just got done with my GATE Exam (Phew!) Onwards to a journey about Learning everything about Natural Language Processing to LLMs with a garnish of Reinforcement Learning (thanks deepseek)

All of it in the following order

NLP Roadmap

• NLP Learning Roadmap
  • Prerequisites
    • Mathematics
      • Linear algebra
      • Probability and statistics
    • Programming
      • Proficiency in a programming language (e.g., Python)
  • Introduction to NLP
    • Definition and scope of NLP
    • Historical development of NLP
    • Key challenges and applications
  • Text Analysis
    • Lexical Analysis
      • Word meaning and structure
      • Morphology (word formation)
      • Lemmatization (base form identification)
    • Syntactic Analysis
      • Parts-of-speech tagging
      • Dependency parsing
      • Constituency parsing
    • Semantic Analysis
      • Extracting meaning
      • Word embedding models (e.g., Word2Vec, GloVe)
      • Topic modeling
    • Additional Semantic Analysis
      • Coreference resolution
      • Discourse analysis
  • Text Processing
    • Tokenization
      • Sentence tokenization
      • Word tokenization
      • Subword tokenization (Byte Pair Encoding, SentencePiece)
    • Stop Words Removal
      • Importance and impact on NLP tasks
      • Customizing stop word lists
    • Stemming and Lemmatization
      • Porter stemming algorithm
      • Snowball stemming algorithm
      • Lemmatization techniques and challenges
    • Part-of-Speech Tagging
      • POS tagging algorithms (HMM-based, rule-based, neural-based)
      • Fine-grained POS tagging
  • Text Representation
    • Bag of Words (BoW)
      • Term Frequency (TF) and Inverse Document Frequency (IDF)
      • Bag of N-grams
    • TF-IDF
      • Calculating TF-IDF scores
      • Applications in information retrieval
    • Word Embeddings
      • Word2Vec
        • Continuous Bag of Words (CBOW) model
        • Skip-gram model
      • GloVe
        • Global Vectors for Word Representation
    • Contextual Embeddings
      • ELMo (Embeddings from Language Models)
      • ULMFiT (Universal Language Model Fine-tuning)
      • OpenAI GPT (Generative Pre-trained Transformer)
  • NLP Libraries and Tools
    • NLTK (Natural Language Toolkit)
    • SpaCy
    • scikit-learn
    • Transformers library (Hugging Face)
  • Statistical Language Models
    • N-grams
      • Unigrams, bigrams, and trigrams
      • N-gram language models
    • Hidden Markov Models (HMM)
      • Basics of HMMs
      • Applications in part-of-speech tagging
  • Machine Learning for NLP
    • Supervised Learning
      • Text classification algorithms (Naive Bayes, Support Vector Machines)
      • Evaluation metrics (precision, recall, F1-score)
    • Named Entity Recognition (NER)
      • Rule-based NER
      • Machine learning-based NER
      • Evaluation metrics for NER
    • Sentiment Analysis
      • Sentiment lexicons
      • Machine learning approaches for sentiment analysis
  • Sequence-to-Sequence Models
    • Recurrent Neural Networks (RNN)
      • Vanishing and exploding gradient problems
      • Bidirectional RNNs
    • Long Short-Term Memory (LSTM)
      • Architecture and key components
      • Gating mechanisms
    • Gated Recurrent Unit (GRU)
      • Simplified gating compared to LSTM
      • Applications and advantages
  • Deep Learning Architectures for NLP
    • Convolutional Neural Networks (CNN) for Text
      • Text classification with CNNs
      • Hierarchical and multi-channel CNNs
    • Transfer Learning in NLP
      • Fine-tuning pre-trained models
      • Universal Sentence Encoder
    • Transformer Architecture
      • Self-attention mechanism
      • Multi-head attention
      • Positional encoding
  • Transduction and Recurrency
    • Transduction in NLP
      • Definition and applications
      • Challenges in sequence-to-sequence transduction
    • Recurrent Neural Networks (RNN)
      • Applications beyond sequence-to-sequence tasks
      • Challenges in training RNNs
  • Advanced Topics in Sequence Modeling
    • Attention Mechanism
      • Scaled Dot-Product Attention
      • Position-wise Feedforward Networks
    • Self-Attention Mechanism
      • The concept of self-attention
      • Layer normalization in self-attention
    • Multi-Head Attention
      • Motivation and benefits
      • Combining multiple attention heads
  • Syntax and Parsing
    • Dependency Parsing
      • Dependency tree representation
      • Transition-based and graph-based parsing
    • Constituency Parsing
      • Treebank representation
      • Earley parsing algorithm
    • Parsing Techniques
      • Chart parsing (CYK parser)
      • Shift-Reduce parsing
  • Semantic Role Labeling (SRL) and Coreference Resolution
    • Semantic Role Labeling
      • PropBank and FrameNet
      • Neural approaches to SRL
    • Coreference Resolution
      • Mention detection
      • End-to-end coreference resolution models
  • Evaluation Metrics
    • Precision, Recall, F1-score
    • BLEU score for machine translation
    • Perplexity for language models
  • NLP in Industry and Research
    • Case studies and applications in various domains (healthcare, finance, legal, etc.)
    • Emerging research trends in NLP
  • Ethical Considerations and Bias in NLP
    • Addressing Bias in NLP Models
      • Identifying and mitigating biases in training data
      • Fairness-aware machine learning
    • Ethical Considerations in NLP Research and Deployment
      • Privacy concerns in NLP
      • Responsible AI practices in NLP
  • Continuous Learning and Keeping Updated
    • Follow conferences (ACL, NAACL, EMNLP)
    • Engage with the NLP community
    • Explore recent research papers and advancements (Arxiv, NeurIPS)
  • Projects and Hands-on Practice
    • Apply knowledge through practical projects
    • Contribute to open-source NLP projects
    • Participate in Kaggle competitions

LLM Roadmap

–Coming Soon–