Online decoding of Markov models under latency constraints

Authors:
Mukund Narasimhan;Paul Viola;Michael Shilman
Affiliations:
Microsoft Corporation, One Microsoft Way, Redmond WA;Microsoft Corporation, One Microsoft Way, Redmond WA;Microsoft Corporation, One Microsoft Way, Redmond WA
Venue:
ICML '06 Proceedings of the 23rd international conference on Machine learning
Year:
2006

Citing 11
Cited 2

A maximum entropy approach to natural language processing

Computational Linguistics
Brief Application Description; Visual Data Mining: Recognizing Telephone Calling Fraud

Data Mining and Knowledge Discovery
Adaptive Fraud Detection

Data Mining and Knowledge Discovery
Distributed Data Mining in Credit Card Fraud Detection

IEEE Intelligent Systems
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data

ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning
Maximum Entropy Markov Models for Information Extraction and Segmentation

ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning
Table extraction using conditional random fields

Proceedings of the 26th annual international ACM SIGIR conference on Research and development in informaion retrieval
Hidden Markov Model Based Continuous Online Gesture Recognition

ICPR '98 Proceedings of the 14th International Conference on Pattern Recognition-Volume 2 - Volume 2
Learning as search optimization: approximate large margin methods for structured prediction

ICML '05 Proceedings of the 22nd international conference on Machine learning
Interactive information extraction with constrained conditional random fields

AAAI'04 Proceedings of the 19th national conference on Artifical intelligence
Discriminative probabilistic models for relational data

UAI'02 Proceedings of the Eighteenth conference on Uncertainty in artificial intelligence

Exploring the Trade-off Between Accuracy and Observational Latency in Action Recognition

International Journal of Computer Vision
Hierarchical multi-channel hidden semi Markov graphical models for activity recognition

Computer Vision and Image Understanding

Quantified Score

Hi-index	0.00

Visualization

Abstract

The Viterbi algorithm is an efficient and optimal method for decoding linear-chain Markov Models. However, the entire input sequence must be observed before the labels for any time step can be generated, and therefore Viterbi cannot be directly applied to online/interactive/streaming scenarios without incurring significant (possibly unbounded) latency. A widely used approach is to break the input stream into fixed-size windows, and apply Viterbi to each window. Larger windows lead to higher accuracy, but result in higher latency.We propose several alternative algorithms to the fixed-sized window decoding approach. These approaches compute a certainty measure on predicted labels that allows us to trade off latency for expected accuracy dynamically, without having to choose a fixed window size up front. Not surprisingly, this more principled approach gives us a substantial improvement over choosing a fixed window. We show the effectiveness of the approach for the task of spotting semi-structured information in large documents. When compared to full Viterbi, the approach suffers a 0.1 percent error degradation with a average latency of 2.6 time steps (versus the potentially infinite latency of Viterbi). When compared to fixed windows Viterbi, we achieve a 40x reduction in error and 6x reduction in latency.