Optimum algorithm to minimize human interactions in sequential Computer Assisted Pattern Recognition

Abstract

Given a Pattern Recognition task, Computer Assisted Pattern Recognition can be viewed as a series of solution proposals made by a computer system, followed by corrections made by a user, until an acceptable solution is found. For this kind of systems, the appropriate measure of performance is the expected number of corrections the user has to make.In the present work we study the special case when the solution proposals have a sequential nature. Some examples of this type of tasks are: language translation, speech transcription and handwriting text transcription. In all these cases the output (the solution proposal) is a sequence of symbols. In this framework it is assumed that the user corrects always the first error found in the proposed solution. As a consequence, the prefix of the proposed solution before the last error correction can be assumed error free in the next iteration.Nowadays, all the techniques in the literature relies in proposing, at each step, the most probable suffix given that a prefix of the "correct" output is already known. Usually the computation of the conditional most probable output is an NP-Hard or an undecidable problem (and then we have to apply some approximations) or, in some simple cases, complex dynamic programming techniques should be used (usualy some variant of the Viterbi algorithm).In the present work we show that this strategy is not optimum when we are interested in minimizing the number of human interactions. Moreover we describe the optimum strategy that is simpler (and usually faster) to compute.

On the optimal decision rule for sequential interactive structured prediction

“…However, this is not optimal for ISP since the decision rule should be formalized in terms of minimizing user interactions. Indeed, this fact was proved in [7], where an alternative strategy is applied to a specific case of ISP (i.e., text prediction). Inspired by that work, here we provide an optimal decision rule for ISP which covers a broader range of common ISP problems in which the output depends on a structured input x.…”

“…if any feature r 4 if any feature is p (1) if any feature iŝ s (1) if any feature is not available at your web y (1) if any feature is not available at your web i = 2 a (2) if any feature is not available r 7 if any feature is not available in p (2) if any feature is not available in s (2) if any feature is not available in your network y (2) if any feature is not available in your network I = 2 y (2) ≡ r if any feature is not available in your network erroneously classified sample by assigning the correct label. This function is known as the zero-one loss function and leads to the maximum-a-posteriori (MAP) decision rule [2]:…”

Recognition of pen-based music notation with finite-state machines

“…The last step is to apply some kind of decoding strategy in order to output a hypothesis concerning the input sequence of strokes. The possibility of approaching this stage following different strategies is another -Optimum decoding to minimize number of corrections (MNC): if it is as-400 sumed that the output will be corrected by a human supervisor, and that after each correction the machine will be allowed to output a new hypothe-sis (interactive approach), the optimum means of minimizing the expected number of sequential corrections is that of computing the algorithm developed by Oncina (2009).…”

Interactive Pattern Recognition applied to Natural Language Processing

“…Surprisingly enough, the maximization of the posterior probability and the entailed Viterbi search is not necesarily the best search strategy in this case. Recently, a better and simpler approach has been proposed [40].…”

“…According to this, in [40] an optimal strategy to predict suffixes in an interactive environment is achieved. This strategy turns out to be a greedy-like search (denoted as Greedy from now on) and it is based on constructing the final hypothesis by taking just optimum local decisions.…”