Speech Recognition

1
Speech Recognition

• Algorithm Aspects in Speech Recognition ,Adam
  L. Buchsbaum ,Raffaele Giancarl
• Presents the main fields of speech recognition
• The general problem areas
• Graph searching
• Automata manipulation
• Shortest path finding
• Finite state automata minimization
• Some of the major open problems from an
  algorithmic viewpoint
• Asymptotically efficient handle very large
  instances
• Practically efficient run in real time

2
Block diagram of speech recognizer
3
IWR , CSR

• IWR Isolated Word Recognition
• Spoken in isolation and belonging to a fixed
  dictionary
• Lexicon typical pronunciations of each word in
  dictionary
• Search algorithm output the word that maximizes
  a given objective function ( likelihood of a word
  given the observation sequence)
• CSR Continuous Speech Recognition
• Lexicon same as IWR
• Language model give a stochastic description of
  the language and the possibly probabilistic
  description of which specific words can follow
  another word or group of words
• Search algorithm find a grammatically correct
  sentence that maximizes a given objective
  function ( likelihood of a sentence given the
  observation sequence)
• Coarticulation effects how to recognize
  speech vs how to wreck a nice beach,
  incomplete information

4
Major methods for speech recognition

• Template-based approach
• Small dictionaries, mainly for IWR
• Reference templates (a sequence of feature
  vectors representing a unit of speech to be
  recognized)
• Distance measure
• eg log spectral distance, likelihood
  distortions
• Stochastic approach (maximum likelihood)
• Dominant method
• Equations represent
• X observation sequence
• W unknown sentence
• Output the sentence w that
• Pr(w) max w Pr(WX)
• Pr(WX) Pr(X) Pr(XW) Pr(W)
• W argmax w Pr(XW) Pr(W)
  for fixed X
• ( argmaxw f(w) w ? f(w)
  max w f(w)
• Defn Cs -logPr eg Cs(W) - logPr(W)
• W argmin w Cs(W) Cs(XW)
• Solution of the equation
• Language Modeling and Acoustic
  Modeling

5
Modeling Tools

• HMM (Hidden Markov Model)
• Quintuple ? (N,M,A,B,p)
• N the number of state
• M the number of symbols that each state can
  output or recognize
• A NN state transition matrix, a(i,j) the
  probability of moving from state i to state j
• B observation probability distribution, bi(d)
  the probability of recognizing or generating the
  symbol dwhen in state i
• ? the initial state probability distribution
  such that ?i the probability of being in state
  i at time 1.
• MS (Markov Source)
• E transitions between states
• V set of states
• ? alphabet including null symbol
• One to one mapping M from E to V?V
• M(t) (i, a, j)
• i is the predecessor state of t
• t output symbol a
• j is the successor state of t

6
Viterbi

• Viterbi Algorithm
• Compute the optimal state sequence Q
  (q1,..,qT) through? that matches X. (that is max
  Pr(QX,?) )
• ßt (i) probability along the highest
  probability that accounts for the first t
  observations and ends in state i
• ?t (i) the state at time t-1 that led to
  state I at time t along that path
• Initialization
• Induction
• Termination
• Backtracking

7
Acoustic Word Models via Acoustic Phone Models

• Tree representation
• Static data structure
• Lexicon
• Over the alphabet of
• feature vectors

8
MS ,HMM

• Circles represent states, arcs represent
  transitions.
• Arcs are labeled f/p, denoting that the
  associated transition outputs phone f and occurs
  with probability p
• For each phone f in the alphabet build a HMM
• Directed graph having a minimum of four and a
  maximum of seven states with exactly one source,
  one sink, self loops and no back arcs
• Gives an acoustic model describing the different
  ways in which one can pronounce the given phone
• Technically, this HMM is a device for computing
  how likely it is that a given observation
  sequence acoustically matches the given phone

9
MS HMM
10
Conclusion

• Language Model
• Pr(W) Pr(w1.wj) Pr(w1)Pr(w2w1)..Pr(wjw1..
  wj-1)
• Approximation Pr(wjw1.wj-1)
  Pr(wjwj-k1.w1)
• 20,000 words,k2,400 million vertices and arcs in
  the model
• Possible solution group the equivalence classes
  (how to divide?)
• Heuristic approach
• Layer solution
• Shortest path finding
• Automata machine
• Redundancy problem and size reduction
• Training with efficiency

11
Application

• ATT Watson Advanced Speech Application Platform
  http//www.att.com/aspg/blasr.html
• BBN Speech Products http//www.bbn.com/speech_prod
  s/
• DragonDictate from Dragon Systems,Inc.
  http//www.dragonsys.com/

12
(No Transcript)