Conclusion

• One of the big problems with language is that lots of information is omitted.
• People assume that other people know it, so they omit it. Machines don't know these assumptions.
• Take Home Points
  1. The full fledged approach involves at least the lexicon, syntax and semantics.
  2. Bag of Word techniques work as an approximation of semantics.
  3. Information extraction, machine translation, conversational systems, and speech recognition are all industrially viable technologies.
  4. Simple NLP is easy, but it gets exponentially harder to become more accurate.
• Reading: For this week is Russell and Norvig NL for Communication chapter (pp. 904-939). A better text is James Allen's Natural Language Understanding.
• Reading: For this week is Russell and Norvig's Adversarial Search chapter through section 4 (pp. 164-179).