Changes between Version 1 and Version 2 of en/WillComputerUnderstand


Ignore:
Timestamp:
Jun 6, 2014, 12:51:43 PM (7 years ago)
Author:
xkocinc
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • en/WillComputerUnderstand

    v1 v2  
    11= Will Computers Ever Understand Us? =
     2
     3[[Image(/trac/research/raw-attachment/wiki/en/WillComputerUnderstand/turing.png)]]
     4
     5== Computer “understanding”: Use cases ==
     6
     7 * inappropriate discussion posts detection
     8 * text summarization
     9 * opinion mining
     10 * content targeting
     11 * question answering
     12
     13
     14== Inappropriate discussion posts detection ==
     15
     16That BOY aint done growing and fcuking so she would be stooopid to tie HERSELF down wit a BABY and a tattoo is just as worse!!!
     17
     18=>
     19
     20That BOY aint done growing and '''fcuking''' so she would be '''stooopid''' to tie HERSELF down wit a BABY and a tattoo is just as worse!!!
     21
     22[[Image(/trac/research/raw-attachment/wiki/en/WillComputerUnderstand/fcuking.png, align=right)]]
     23
     24'''Use case:''' discussion forum, automatic detection of inappropriate posts
     25
     26'''Common solution:''' word list
     27
     28'''But:''' users use concealed words that are difficult to detect (f*king, f.u.c.k, f..k, fcuking)
     29
     30'''Better solution:''' word list + concealing rules
     31
     32'''But:''' users invent new words and concealing patterns
     33
     34'''Even better solution:''' word list + automatically generated thesaurus + concealing rules + metarules
     35
     36
     37
     38== Text summarization ==
     39
     40[[Image(/trac/research/raw-attachment/wiki/en/WillComputerUnderstand/text_sum.png)]]
     41
     42A decade ago a girl spent money for ringtones.
     43
     44
     45'''Use case:''' automatic abstract generation, multiple document digest, are these documents stating similar or opposite arguments?
     46
     47'''Naive solution:''' take every first sentence in a paragraph
     48
     49'''Common solution:''' take every sentence containing a keyword
     50
     51'''But:''' not really scalable, difficult to detect the main message
     52
     53'''Better approach:'''
     54 1. analyse text on several levels
     55   * whole document (sections, paragraphs, consistency)
     56   * sequence of sentences (each having a structure)
     57   * bag of words and keywords (in different forms, synonyms, abbreviations etc.)
     58 1. generate a summary
     59
     60
     61== Opinion mining ==
     62
     63The ''iPhone 5'' price was predictably high and continues to be so, so consumers will need to bear that in mind too when looking for their next smartphone.
     64
     65...
     66
     67Well, all of those picking up the iPhone 5 will have the same reaction: this thing is amazingly light. You’ve probably heard the
     68numbers by now (20 per cent lighter than the predecessor, as well as beating most of the opposition too at 112g.)
     69
     70
     71=>
     72
     73
     74The ''iPhone 5'' '''price''' was predictably '''high''' and continues to be so, so consumers will need to bear that in mind too when looking for their next smartphone.
     75
     76...
     77
     78Well, all of those picking up the '''iPhone 5''' will have the same reaction: this thing is '''amazingly light'''. You’ve probably heard the numbers by now ('''20 per cent lighter''' than the predecessor, as well as beating most of the opposition too at 112g.)
     79
     80
     81[[Image(/trac/research/raw-attachment/wiki/en/WillComputerUnderstand/okko.png)]]
     82
     83
     84'''Use case:''' what are people thinking about a particular product/company/idea X?
     85
     86'''Solution:''' search X, find evaluative words
     87
     88'''But:''' opinions are expressed by non-evaluative words
     89
     90'''Better solution:'''
     91 * extract useful attributes of X (noise, weight, price, appearance)
     92 * generate thesauri of evaluative words: thin iPhone 5 × thin tasteless burger
     93
     94
     95== Question answering ==
     96
     97Do you have a bike for a 4-year-old girl?
     98'''Search results for “bike”, “girl”'''
     99
     100...
     101
     102---
     103
     104Do you have a bike for a 4-year-old girl?
     105
     106'''If she is under 110 cm tall I will recommend Maggie, Princess or Misty. If she is taller I would recommend Miss B or Kellie. If she does not insist on bike for girls I would also recommend Racer or Mr. Lightning. How tall is she?'''
     107
     108About 105 cm.
     109
     110'''Do you have some other constraints?'''
     111
     112I look for something cheaper.
     113
     114'''Then I would recommend Princess. It is a popular bike.'''
     115
     116---
     117
     118'''Use case:''' chatbot providing basic support
     119
     120'''Solution:''' patterns, keyword detection, searching
     121
     122'''But:''' no real dialogue, no real answers, just searching
     123
     124'''Better solution:''' sentence structure analysis, keyword detection, coreference resolution, dialogue strategy
     125
     126[[Image(/trac/research/raw-attachment/wiki/en/WillComputerUnderstand/bike.png)]]
     127
     128
     129== Conclusions: Understanding of ''understanding'' ==
     130
     131Is this real understanding?
     132
     133Probably not.
     134
     135We do not know what understanding is but we know how it looks like when someone understands.
     136
     137Computer programs that can discover a vulgar text, summarize a text, recognize someone’s feelings or answer questions
     138look like they understand our language...