Lightweight Random Indexing for Polylingual Text Classification

Abstract

Multilingual Text Classification (MLTC) is a text classification task in which documents are written each in one among a set L of natural languages, and in which all documents must be classified under the same classification scheme, irrespective of language. There are two main variants of MLTC, namely Cross-Lingual Text Classification (CLTC) and Polylingual Text Classification (PLTC). In PLTC, which is the focus of this paper, we assume (differently from CLTC) that for each language in L there is a representative set of training documents; PLTC consists of improving the accuracy of each of the |L| monolingual classifiers by also leveraging the training documents written in the other (|L| − 1) languages. The obvious solution, consisting of generating a single polylingual classifier from the juxtaposed monolingual vector spaces, is usually infeasible, since the dimensionality of the resulting vector space is roughly |L| times that of a monolingual one, and is thus often unmanageable. As a response, the use of machine translation tools or multilingual dictionaries has been proposed. However, these resources are not always available, or are not always free to use.One machine-translation-free and dictionary-free method that, to the best of our knowledge, has never been applied to PLTC before, is Random Indexing (RI). We analyse RI in terms of space and time efficiency, and propose a particular configuration of it (that we dub Lightweight Random Indexing -LRI). By running experiments on two well known public benchmarks, Reuters RCV1/RCV2 (a comparable corpus) and JRC-Acquis (a parallel one), we show LRI to outperform (both in terms of effectiveness and efficiency) a number of previously proposed machine-translation-free and dictionary-free PLTC methods that we use as baselines.

Identifying constitutive articles of cumulative dissertation theses by bilingual text similarity. Evaluation of similarity methods on a new short text task

“…Once we look at the ranking results on the level of matching to individual theses, the limitations of the VSM become apparent as the Random Indexing method performs clearly better. The multilingual application of RI has so far only received limited attention (Fernández, Esuli, & Sebastiani, 2016;Moen & Marsi, 2013;Sahlgren & Karlgren, 2005) but the present results are very encouraging. The findings also indicate that the trigram and fastText methods perform moderately well while LSA is not competitive for this particular task.…”

Heterogeneous document embeddings for cross-lingual text classification

“…We use the same learner as in [3], i.e., Support Vector Machines (SVMs), as implemented in the scikit-learn package. 5 For the 2nd-tier classifier of gFun, and for all the baseline methods, we optimize the C parameter, that trades off between training error and margin, testing all values of C = 10 i for i ∈ {−1, ..., 4} via kfold cross-validation. We use Platt calibration in order to calibrate the 1st-tier classifiers.…”

“…Naïve LRI [5] CLESA [7] KCCA [9] DCI [4] Fun [3] Baselines. As the baselines against which to compare gFun we use the naïve monolingual baseline (hereafter indicated as Naïve), Funnelling (Fun), plus the four best baselines of [3], namely, Lightweight Random Indexing (LRI) [5], Cross-Lingual Explicit Semantic Analysis (CLESA) [7], Kernel Canonical Correlation Analysis (KCCA) [9], and Distributional Correspondence Indexing (DCI) [4]. For all systems but gFun, the results we report are excerpted from [3], so we refer to that paper for the detailed setups of these baselines.…”

Divide-and-Learn: A Random Indexing Approach to Attribute Inference Attacks in Online Social Networks

“…As for RI, [7] proposes a grid of sample values. We set dimension= 500 and select two non-zero elements of the index vector to {-1, +1}, which maximize the result of our inference attacks.…”