Page 371 - AI_Ver_3.0_class_11
P. 371
3. Learning: Learn and understand the meaning of facts related to the topic. Gather and arrange facts supporting your
argument and formulate your arrangement each time when new evidence arrives. This will help you find updated or
completely new information that can score points against your opponent’s position.
4. Interacting: Listening to your opponent’s arguments and opinions, then present a convincing altercation that
further proves your case.
Understanding Reasoning Learning Interacting
Cognitive systems They reason underlying They never stop learning. They interact with
understand like humans do. ideas and concepts. They They develop "expertise" humans.
debate, infer, and extract with every interaction and
concepts. outcome.
Task #Creativity
Find five interesting facts about “IBM Project Debater” and write in the space provided below. [CBSE Handbook]
Natural Language Processing (NLP)
Computers excel in working with structured data, in which everything is properly organised and labelled. Unfortunately
for machines, human language is not structured. You've spent most of your life communicating through human language.
Your brain accomplishes this with some of the most complex neural circuitry on this planet. However, creating machines
that can understand human language, is extremely challenging.
NLP involves machines segmenting sentences and extracting meaning from “tokens” of human language. Human
language is difficult and unstructured. Despite being loosely held together by grammar constraints; our language
presents information in a variety of puzzling ways. Unstructured information, unlike structured information that can
be organised in tables or matrices with clearly labelled rows and columns, is disorganised and difficult to comprehend.
Consider the statements made by famous personalities:
• Groucho Marx: “One morning I shot an elephant in my pajamas. How he got into my pajamas I'll never know.”
• Steven Wright: “I poured spot remover on my dog. Now he’s gone.”
• George Carlin: “I went to a bookstore and asked the saleswoman, ‘Where’s the self-help section?’ She said if she
told me, it would defeat the purpose.”
To deal with the “messiness” of unstructured data, computers start with one sentence at a time. The process is known
as sentence segmentation that is sentence-based text analysis. In NLP analysis, the text data is either analysed using
meaningful words called tokens, or analysed using sentences. Computers divide the input into small piece of information
(tokens), which may be classified separately. NLP can deal with text tokens after they have been sorted into a structured
format or manner based on their meaning.
Leveraging Linguistics and Computer Science 369

