There are 293 scripts representing over 7,000 languages in the written form. Due to various reasons, many closely related languages use different scripts, which poses difficulty for multilingual pretrained language models (mPLMs) in learning crosslingual knowledge through lexical overlap. As a result, mPLMs present a script barrier: representations from different scripts are located in different subspaces, which is a strong indicator of why crosslingual transfer involving languages of different scripts shows sub-optimal performance. To address this problem, we propose a simple framework TransliCo that contains Transliteration Contrastive Modeling (TCM) to fine-tune an mPLM by contrasting sentences in its training data and their transliterations in a unified script (Latn, in our case), which ensures uniformity in the representation space for different scripts. Using Glot500-m, an mPLM pretrained on over 500 languages, as our source model, we find-tune it on a small portion (5\%) of its training data, and refer to the resulting model as Furina. We show that Furina not only better aligns representations from distinct scripts but also outperforms the original Glot500-m on various crosslingual transfer tasks. Additionally, we achieve consistent improvement in a case study on the Indic group where the languages are highly related but use different scripts. We make our code and models publicly available.
Transformer-based pre-trained language models (PLMs) have achieved remarkable performance in various natural language processing (NLP) tasks. However, pre-training such models can take considerable resources that are almost only available to high-resource languages. On the contrary, static word embeddings are easier to train in terms of computing resources and the amount of data required. In this paper, we introduce MoSECroT Model Stitching with Static Word Embeddings for Crosslingual Zero-shot Transfer), a novel and challenging task that is especially relevant to low-resource languages for which static word embeddings are available. To tackle the task, we present the first framework that leverages relative representations to construct a common space for the embeddings of a source language PLM and the static word embeddings of a target language. In this way, we can train the PLM on source-language training data and perform zero-shot transfer to the target language by simply swapping the embedding layer. However, through extensive experiments on two classification datasets, we show that although our proposed framework is competitive with weak baselines when addressing MoSECroT, it fails to achieve competitive results compared with some strong baselines. In this paper, we attempt to explain this negative result and provide several thoughts on possible improvement.
Pretraining multilingual language models from scratch requires considerable computational resources and substantial training data. Therefore, a more efficient method is to adapt existing pretrained language models (PLMs) to new languages via vocabulary extension and continued pretraining. However, this method usually randomly initializes the embeddings of new subwords and introduces substantially more embedding parameters to the language model, thus weakening the efficiency. To address these issues, we propose a novel framework: \textbf{O}ne \textbf{F}or \textbf{A}ll (\textbf{\textsc{Ofa}}), which wisely initializes the embeddings of unseen subwords from target languages and thus can adapt a PLM to multiple languages efficiently and effectively. \textsc{Ofa} takes advantage of external well-aligned multilingual word embeddings and injects the alignment knowledge into the new embeddings. In addition, \textsc{Ofa} applies matrix factorization and replaces the cumbersome embeddings with two lower-dimensional matrices, which significantly reduces the number of parameters while not sacrificing the performance. Through extensive experiments, we show models initialized by \textsc{Ofa} are efficient and outperform several baselines. \textsc{Ofa} not only accelerates the convergence of continued pretraining, which is friendly to a limited computation budget, but also improves the zero-shot crosslingual transfer on a wide range of downstream tasks. We make our code and models publicly available.
Although unsupervised neural machine translation (UNMT) has achieved success in many language pairs, the copying problem, i.e., directly copying some parts of the input sentence as the translation, is common among distant language pairs, especially when low-resource languages are involved. We find this issue is closely related to an unexpected copying behavior during online back-translation (BT). In this work, we propose a simple but effective training schedule that incorporates a language discriminator loss. The loss imposes constraints on the intermediate translation so that the translation is in the desired language. By conducting extensive experiments on different language pairs, including similar and distant, high and low-resource languages, we find that our method alleviates the copying problem, thus improving the translation performance on low-resource languages.
An interesting line of research in natural language processing (NLP) aims to incorporate linguistic typology to bridge linguistic diversity and assist the research of low-resource languages. While most works construct linguistic similarity measures based on lexical or typological features, such as word order and verbal inflection, recent work has introduced a novel approach to defining language similarity based on how they represent basic concepts, which is complementary to existing similarity measures. In this work, we study the conceptual similarity in detail and evaluate it extensively on a binary classification task.
Colexification in comparative linguistics refers to the phenomenon of a lexical form conveying two or more distinct meanings. In this paper, we propose simple and effective methods to build multilingual graphs from colexification patterns: ColexNet and ColexNet+. ColexNet's nodes are concepts and its edges are colexifications. In ColexNet+, concept nodes are in addition linked through intermediate nodes, each representing an ngram in one of 1,334 languages. We use ColexNet+ to train high-quality multilingual embeddings $\overrightarrow{\mbox{ColexNet+}}$ that are well-suited for transfer learning scenarios. Existing work on colexification patterns relies on annotated word lists. This limits scalability and usefulness in NLP. In contrast, we identify colexification patterns of more than 2,000 concepts across 1,335 languages directly from an unannotated parallel corpus. In our experiments, we first show that ColexNet has a high recall on CLICS, a dataset of crosslingual colexifications. We then evaluate $\overrightarrow{\mbox{ColexNet+}}$ on roundtrip translation, verse retrieval and verse classification and show that our embeddings surpass several baselines in a transfer learning setting. This demonstrates the benefits of colexification for multilingual NLP.
Languages differ in how they divide up the world into concepts and words; e.g., in contrast to English, Swahili has a single concept for `belly' and `womb'. We investigate these differences in conceptualization across 1,335 languages by aligning concepts in a parallel corpus. To this end, we propose Conceptualizer, a method that creates a bipartite directed alignment graph between source language concepts and sets of target language strings. In a detailed linguistic analysis across all languages for one concept (`bird') and an evaluation on gold standard data for 32 Swadesh concepts, we show that Conceptualizer has good alignment accuracy. We demonstrate the potential of research on conceptualization in NLP with two experiments. (1) We define crosslingual stability of a concept as the degree to which it has 1-1 correspondences across languages, and show that concreteness predicts stability. (2) We represent each language by its conceptualization pattern for 83 concepts, and define a similarity measure on these representations. The resulting measure for the conceptual similarity of two languages is complementary to standard genealogical, typological, and surface similarity measures. For four out of six language families, we can assign languages to their correct family based on conceptual similarity with accuracy between 54\% and 87\%.
In this paper, we model, analyze and optimize the multi-user and multi-order-reflection (MUMOR) intelligent reflecting surface (IRS) networks. We first derive a complete MUMOR IRS network model applicable for the arbitrary times of reflections, size and number of IRSs/reflectors. The optimal condition for achieving sum-rate upper bound with one IRS in a closed-form function and the analytical condition to achieve interference-free transmission are derived, respectively. Leveraging this optimal condition, we obtain the MUMOR sum-rate upper bound of the IRS network with different network topologies, where the linear graph (LG), complete graph (CG) and null graph (NG) topologies are considered. Simulation results verify our theories and derivations and demonstrate that the sum-rate upper bounds of different network topologies are under a K-fold improvement given K-piece IRS.
In this work, we propose a flow-adapter architecture for unsupervised NMT. It leverages normalizing flows to explicitly model the distributions of sentence-level latent representations, which are subsequently used in conjunction with the attention mechanism for the translation task. The primary novelties of our model are: (a) capturing language-specific sentence representations separately for each language using normalizing flows and (b) using a simple transformation of these latent representations for translating from one language to another. This architecture allows for unsupervised training of each language independently. While there is prior work on latent variables for supervised MT, to the best of our knowledge, this is the first work that uses latent variables and normalizing flows for unsupervised MT. We obtain competitive results on several unsupervised MT benchmarks.
Secret key generation in physical layer security exploits the unpredictable random nature of wireless channels. However, the millimeter wave (mmWave) channels have limited multipath and may not be Gaussian distributed. In this paper, for mmWave secret key generation of physical layer security, we use intelligent reflecting surface (IRS) to produce randomness and induce artificial Rayleigh fading directly in the wireless environments. We first formulate the model of IRS-assisted key generation in mmWave environments. The IRS-assisted reflection channel varies according to the IRS weights' variation and induces randomness. When considering the IRS weights are continuous and discrete uniformly distributed, we find that the reflection channel variance is equal to the number of IRS elements. Besides, we prove that the magnitude and phase are Rayleigh and uniformly distributed when the weights are continuously and discretely distributed with more than one quantization bit. With the simulation results verifying the analytical results, this work explains the mathematical principles behind the IRS-assisted physical layer security secret key generation and lays a foundation for future mmWave key generation evaluation and optimization of channel randomness.