Model-agnostic meta-learning (MAML) has been recently put forth as a strategy to learn resource-poor languages in a sample-efficient fashion. Nevertheless, the properties of these languages are often not well represented by those available during training. Hence, we argue that the i.i.d. assumption ingrained in MAML makes it ill-suited for cross-lingual NLP. In fact, under a decision-theoretic framework, MAML can be interpreted as minimising the expected risk across training languages (with a uniform prior), which is known as Bayes criterion. To increase its robustness to outlier languages, we create two variants of MAML based on alternative criteria: Minimax MAML reduces the maximum risk across languages, while Neyman-Pearson MAML constrains the risk in each language to a maximum threshold. Both criteria constitute fully differentiable two-player games. In light of this, we propose a new adaptive optimiser solving for a local approximation to their Nash equilibrium. We evaluate both model variants on two popular NLP tasks, part-of-speech tagging and question answering. We report gains for their average and minimum performance across low-resource languages in zero- and few-shot settings, compared to joint multi-source transfer and vanilla MAML.
The goal of generative phonology, as formulated by Chomsky and Halle (1968), is to specify a formal system that explains the set of attested phonological strings in a language. Traditionally, a collection of rules (or constraints, in the case of optimality theory) and underlying forms (UF) are posited to work in tandem to generate phonological strings. However, the degree of abstraction of UFs with respect to their concrete realizations is contentious. As the main contribution of our work, we implement the phonological generative system as a neural model differentiable end-to-end, rather than as a set of rules or constraints. Contrary to traditional phonology, in our model, UFs are continuous vectors in $\mathbb{R}^d$, rather than discrete strings. As a consequence, UFs are discovered automatically rather than posited by linguists, and the model can scale to the size of a realistic vocabulary. Moreover, we compare several modes of the generative process, contemplating: i) the presence or absence of an underlying representation in between morphemes and surface forms (SFs); and ii) the conditional dependence or independence of UFs with respect to SFs. We evaluate the ability of each mode to predict attested phonological strings on 2 datasets covering 5 and 28 languages, respectively. The results corroborate two tenets of generative phonology, viz. the necessity for UFs and their independence from SFs. In general, our neural model of generative phonology learns both UFs and SFs automatically and on a large-scale.
The success of large pretrained language models (LMs) such as BERT and RoBERTa has sparked interest in probing their representations, in order to unveil what types of knowledge they implicitly capture. While prior research focused on morphosyntactic, semantic, and world knowledge, it remains unclear to which extent LMs also derive lexical type-level knowledge from words in context. In this work, we present a systematic empirical analysis across six typologically diverse languages and five different lexical tasks, addressing the following questions: 1) How do different lexical knowledge extraction strategies (monolingual versus multilingual source LM, out-of-context versus in-context encoding, inclusion of special tokens, and layer-wise averaging) impact performance? How consistent are the observed effects across tasks and languages? 2) Is lexical knowledge stored in few parameters, or is it scattered throughout the network? 3) How do these representations fare against traditional static word vectors in lexical tasks? 4) Does the lexical information emerging from independently trained monolingual LMs display latent similarities? Our main results indicate patterns and best practices that hold universally, but also point to prominent variations across languages and tasks. Moreover, we validate the claim that lower Transformer layers carry more type-level lexical knowledge, but also show that this knowledge is distributed across multiple layers.
In order to simulate human language capacity, natural language processing systems must complement the explicit information derived from raw text with the ability to reason about the possible causes and outcomes of everyday situations. Moreover, the acquired world knowledge should generalise to new languages, modulo cultural differences. Advances in machine commonsense reasoning and cross-lingual transfer depend on the availability of challenging evaluation benchmarks. Motivated by both demands, we introduce Cross-lingual Choice of Plausible Alternatives (XCOPA), a typologically diverse multilingual dataset for causal commonsense reasoning in 11 languages. We benchmark a range of state-of-the-art models on this novel dataset, revealing that current methods based on multilingual pretraining and zero-shot fine-tuning transfer suffer from the curse of multilinguality and fall short of performance in monolingual settings by a large margin. Finally, we propose ways to adapt these models to out-of-sample resource-lean languages where only a small corpus or a bilingual dictionary is available, and report substantial improvements over the random baseline. XCOPA is available at github.com/cambridgeltl/xcopa.
In previous work, artificial agents were shown to achieve almost perfect accuracy in referential games where they have to communicate to identify images. Nevertheless, the resulting communication protocols rarely display salient features of natural languages, such as compositionality. In this paper, we propose some realistic sources of pressure on communication that avert this outcome. More specifically, we formalise the principle of least effort through an auxiliary objective. Moreover, we explore several game variants, inspired by the principle of object constancy, in which we alter the frequency, position, and luminosity of the objects in the images. We perform an extensive analysis on their effect through compositionality metrics, diagnostic classifiers, and zero-shot evaluation. Our findings reveal that the proposed sources of pressure result in emerging languages with less redundancy, more focus on high-level conceptual information, and better abilities of generalisation. Overall, our contributions reduce the gap between emergent and natural languages.
We introduce Multi-SimLex, a large-scale lexical resource and evaluation benchmark covering datasets for 12 typologically diverse languages, including major languages (e.g., Mandarin Chinese, Spanish, Russian) as well as less-resourced ones (e.g., Welsh, Kiswahili). Each language dataset is annotated for the lexical relation of semantic similarity and contains 1,888 semantically aligned concept pairs, providing a representative coverage of word classes (nouns, verbs, adjectives, adverbs), frequency ranks, similarity intervals, lexical fields, and concreteness levels. Additionally, owing to the alignment of concepts across languages, we provide a suite of 66 cross-lingual semantic similarity datasets. Due to its extensive size and language coverage, Multi-SimLex provides entirely novel opportunities for experimental evaluation and analysis. On its monolingual and cross-lingual benchmarks, we evaluate and analyze a wide array of recent state-of-the-art monolingual and cross-lingual representation models, including static and contextualized word embeddings (such as fastText, M-BERT and XLM), externally informed lexical representations, as well as fully unsupervised and (weakly) supervised cross-lingual word embeddings. We also present a step-by-step dataset creation protocol for creating consistent, Multi-Simlex-style resources for additional languages. We make these contributions -- the public release of Multi-SimLex datasets, their creation protocol, strong baseline results, and in-depth analyses which can be be helpful in guiding future developments in multilingual lexical semantics and representation learning -- available via a website which will encourage community effort in further expansion of Multi-Simlex to many more languages. Such a large-scale semantic resource could inspire significant further advances in NLP across languages.
Unsupervised pretraining models have been shown to facilitate a wide range of downstream applications. These models, however, still encode only the distributional knowledge, incorporated through language modeling objectives. In this work, we complement the encoded distributional knowledge with external lexical knowledge. We generalize the recently proposed (state-of-the-art) unsupervised pretraining model BERT to a multi-task learning setting: we couple BERT's masked language modeling and next sentence prediction objectives with the auxiliary binary word relation classification, through which we inject clean linguistic knowledge into the model. Our initial experiments suggest that our "linguistically-informed" BERT (LIBERT) yields performance gains over the linguistically-blind "vanilla" BERT on several language understanding tasks.
Semantic specialization is the process of fine-tuning pre-trained distributional word vectors using external lexical knowledge (e.g., WordNet) to accentuate a particular semantic relation in the specialized vector space. While post-processing specialization methods are applicable to arbitrary distributional vectors, they are limited to updating only the vectors of words occurring in external lexicons (i.e., seen words), leaving the vectors of all other words unchanged. We propose a novel approach to specializing the full distributional vocabulary. Our adversarial post-specialization method propagates the external lexical knowledge to the full distributional space. We exploit words seen in the resources as training examples for learning a global specialization function. This function is learned by combining a standard L2-distance loss with an adversarial loss: the adversarial component produces more realistic output vectors. We show the effectiveness and robustness of the proposed method across three languages and on three tasks: word similarity, dialog state tracking, and lexical simplification. We report consistent improvements over distributional word vectors and vectors specialized by other state-of-the-art specialization frameworks. Finally, we also propose a cross-lingual transfer method for zero-shot specialization which successfully specializes a full target distributional space without any lexical knowledge in the target language and without any bilingual data.
Addressing the cross-lingual variation of grammatical structures and meaning categorization is a key challenge for multilingual Natural Language Processing. The lack of resources for the majority of the world's languages makes supervised learning not viable. Moreover, the performance of most algorithms is hampered by language-specific biases and the neglect of informative multilingual data. The discipline of Linguistic Typology provides a principled framework to compare languages systematically and empirically and documents their variation in publicly available databases. These enshrine crucial information to design language-independent algorithms and refine techniques devised to mitigate the above-mentioned issues, including cross-lingual transfer and multilingual joint models, with typological features. In this survey, we demonstrate that typology is beneficial to several NLP applications, involving both semantic and syntactic tasks. Moreover, we outline several techniques to extract features from databases or acquire them automatically: these features can be subsequently integrated into multilingual models to tie parameters together cross-lingually or gear a model towards a specific language. Finally, we advocate for a new typology that accounts for the patterns within individual examples rather than entire languages, and for graded categories rather than discrete ones, in oder to bridge the gap with the contextual and continuous nature of machine learning algorithms.
Distributed representations of sentences have been developed recently to represent their meaning as real-valued vectors. However, it is not clear how much information such representations retain about the polarity of sentences. To study this question, we decode sentiment from unsupervised sentence representations learned with different architectures (sensitive to the order of words, the order of sentences, or none) in 9 typologically diverse languages. Sentiment results from the (recursive) composition of lexical items and grammatical strategies such as negation and concession. The results are manifold: we show that there is no `one-size-fits-all' representation architecture outperforming the others across the board. Rather, the top-ranking architectures depend on the language and data at hand. Moreover, we find that in several cases the additive composition model based on skip-gram word vectors may surpass supervised state-of-art architectures such as bidirectional LSTMs. Finally, we provide a possible explanation of the observed variation based on the type of negative constructions in each language.