We propose a succinct randomized language model which employs a perfect hash function to encode fingerprints of n-grams and their associated probabilities, backoff weights, or other parameters. The scheme can represent any standard n-gram model and is easily combined with existing model reduction techniques such as entropy-pruning. We demonstrate the space-savings of the scheme via machine translation experiments within a distributed language modeling framework. 1

This paper explores the use of clickthrough data for query spelling correction. First, large amounts of query-correction pairs are derived by analyzing users ' query reformulation behavior encoded in the clickthrough data. Then, a phrase-based error model that accounts for the transformation probability between multi-term phrases is trained and integrated into a query speller system. Experiments are carried out on a human-labeled data set. Results show that the system using the phrase-based error model outperforms significantly its baseline systems. 1

It has been widely observed that search queries are composed in a very different style from that of the body or the title of a document. Many techniques explicitly accounting for this language style discrepancy have shown promising results for information retrieval, yet a large scale analysis on the extent of the language differences has been lacking. In this paper, we present an extensive study on this issue by examining the language model properties of search queries and the three text streams associated with each web document: the body, the title, and the anchor text. Our information theoretical analysis shows that queries seem to be composed in a way most similar to how authors summarize documents in anchor texts or titles, offering a quantitative explanation to the observations in past work. We apply these web scale n-gram language models to three search query processing (SQP) tasks: query spelling correction, query bracketing and long query segmentation. By controlling the size and the order of different language models, we find that the perplexity metric to be a good accuracy indicator for these query processing tasks. We show that using smoothed language models yields significant accuracy gains for query bracketing for instance, compared to using web counts as in the literature. We also demonstrate that applying web-scale language models can have marked accuracy advantage over smaller ones.

This paper makes three significant extensions to a noisy channel speller designed for standard written text to target the challenging domain of search queries. First, the noisy channel model is subsumed by a more general ranker, which allows a variety of features to be easily incorporated. Second, a distributed infrastructure is proposed for training and applying Web scale n-gram language models. Third, a new phrase-based error model is presented. This model places a probability distribution over transformations between multi-word phrases, and is estimated using large amounts of query-correction pairs derived from search logs. Experiments show that each of these extensions leads to significant improvements over the state-of-the-art baseline methods. 1

In this paper, we systematically assess the value of using web-scale N-gram data in state-of-the-art supervised NLP classifiers. We compare classifiers that include or exclude features for the counts of various N-grams, where the counts are obtained from a web-scale auxiliary corpus. We show that including N-gram count features can advance the state-of-the-art accuracy on standard data sets for adjective ordering, spelling correction, noun compound bracketing, and verb part-of-speech disambiguation. More importantly, when operating on new domains, or when labeled training data is not plentiful, we show that using web-scale N-gram features is essential for achieving robust performance.

N-gram language models are a major resource bottleneck in machine translation. In this paper, we present several language model implementations that are both highly compact and fast to query. Our fastest implementation is as fast as the widely used SRILM while requiring only 25 % of the storage. Our most compact representation can store all 4 billion n-grams and associated counts for the Google n-gram corpus in 23 bits per n-gram, the most compact lossless representation to date, and even more compact than recent lossy compression techniques. We also discuss techniques for improving query speed during decoding, including a simple but novel language model caching technique that improves the query speed of our language models (and SRILM) by up to 300%. 1

We present Tightly Packed Tries (TPTs), a compact implementation of read-only, compressed trie structures with fast on-demand paging and short load times. We demonstrate the benefits of TPTs for storing n-gram back-off language models and phrase tables for statistical machine translation. Encoded as TPTs, these databases require less space than flat text file representations of the same data compressed with the gzip utility. At the same time, they can be mapped into memory quickly and be searched directly in time linear in the length of the key, without the need to decompress the entire file. The overhead for local decompression during search is marginal. 1

We present a simple technique for learning better SVMs using fewer training examples. Rather than using the standard SVM regularization, we regularize toward low weight-variance. Our new SVM objective remains a convex quadratic function of the weights, and is therefore computationally no harder to optimize than a standard SVM. Variance regularization is shown to enable dramatic improvements in the learning rates of SVMs on three lexical disambiguation tasks. 1

Efficient processing of tera-scale text data is an important research topic. This paper proposes lossless compression of N-gram language models based on LOUDS, a succinct data structure. LOUDS succinctly represents a trie with M nodes as a 2M +1bit string. We compress it further for the N-gram language model structure. We also use ‘variable length coding ’ and ‘block-wise compression ’ to compress values associated with nodes. Experimental results for three large-scale N-gram compression tasks achieved a significant compression rate without any loss. 1

Abstract not found