We present the e-Llama models: 8 billion and 70 billion parameter large language models that are adapted towards the e-commerce domain.These models are meant as foundation models with deep knowledge about e-commerce, that form a base for instruction- and fine-tuning.The e-Llama models are obtained by continuously pretraining the Llama 3.1 base models on 1 trillion tokens of domain-specific data.We discuss our approach and motivate our choice of hyperparameters with a series of ablation studies.To quantify how well the models have been adapted to the e-commerce domain, we define and implement a set of multilingual, e-commerce specific evaluation tasks.We show that, when carefully choosing the training setup, the Llama 3.1 models can be adapted towards the new domain without sacrificing significant performance on general domain tasks.We also explore the possibility of merging the adapted model and the base model for a better control of the performance trade-off between domains.

We present LINGUIST, a method for generating annotated data for Intent Classification and Slot Tagging (IC+ST), via fine-tuning AlexaTM 5B, a 5-billion-parameter multilingual sequence-to-sequence (seq2seq) model, on a flexible instruction prompt. In a 10-shot novel intent setting for the SNIPS dataset, LINGUIST surpasses state-of-the-art approaches (Back-Translation and Example Extrapolation) by a wide margin, showing absolute improvement for the target intents of +1.9 points on IC Recall and +2.5 points on ST F1 Score. In the zero-shot cross-lingual setting of the mATIS++ dataset, LINGUIST out-performs a strong baseline of Machine Translation with Slot Alignment by +4.14 points absolute on ST F1 Score across 6 languages, while matching performance on IC. Finally, we verify our results on an internal large-scale multilingual dataset for conversational agent IC+ST and show significant improvements over a baseline which uses Back-Translation, Paraphrasing and Slot Catalog Resampling. To our knowledge, we are the first to demonstrate instruction fine-tuning of a large-scale seq2seq model to control the outputs of multilingual intent- and slot-labeled data generation.

Scaling conversational personal assistants to a multitude of languages puts high demands on collecting and labelling data, a setting in which cross-lingual learning techniques can help to reconcile the need for well-performing Natural Language Understanding (NLU) with a desideratum to support many languages without incurring unacceptable cost. In this work, we show that automatically annotating unlabeled utterances using Machine Translation in an offline fashion and adding them to the training data can improve performance for existing NLU features for low-resource languages, where a straightforward translate-test approach as considered in existing literature would fail the latency requirements of a live environment. We demonstrate the effectiveness of our method with intrinsic and extrinsic evaluation using a real-world commercial dialog system in German. Beyond an intrinsic evaluation, where 56% of the resulting automatically labeled utterances had a perfect match with ground-truth labels, we see significant performance improvements in an extrinsic evaluation settings when manual labeled data is available in small quantities.

We describe the design, the setup, and the evaluation results of the DiscoMT 2017 shared task on cross-lingual pronoun prediction. The task asked participants to predict a target-language pronoun given a source-language pronoun in the context of a sentence. We further provided a lemmatized target-language human-authored translation of the source sentence, and automatic word alignments between the source sentence words and the target-language lemmata. The aim of the task was to predict, for each target-language pronoun placeholder, the word that should replace it from a small, closed set of classes, using any type of information that can be extracted from the entire document. We offered four subtasks, each for a different language pair and translation direction: English-to-French, English-to-German, German-to-English, and Spanish-to-English. Five teams participated in the shared task, making submissions for all language pairs. The evaluation results show that most participating teams outperformed two strong n-gram-based language model-based baseline systems by a sizable margin.

To ensure portability of NLP systems across multiple domains, existing treebanks are often extended by adding trees from interesting domains that were not part of the initial annotation effort. In this paper, we will argue that it is both useful from an application viewpoint and enlightening from a linguistic viewpoint to detect and reduce divergence in annotation schemes between extant and new parts in a set of treebanks that is to be used in evaluation experiments. The results of our correction and harmonization efforts will be made available to the public as a test suite for the evaluation of constituent parsing.

The Live Memories corpus is an Italian corpus annotated for anaphoric relations. This annotation effort aims to contribute to two significant issues for the CL research: the lack of annotated anaphoric resources for Italian and the increasing interest for the social Web. The Live Memories Corpus contains texts from the Italian Wikipedia about the region Trentino/Süd Tirol and from blog sites with users' comments. It is planned to add a set of articles of local news papers. The corpus includes manual annotated information about morphosyntactic agreement, anaphoricity, and semantic class of the NPs. The anaphoric annotation includes discourse deixis, bridging relations and markes cases of ambiguity with the annotation of alternative interpretations. For the annotation of the anaphoric links the corpus takes into account specific phenomena of the Italian language like incorporated clitics and phonetically non realized pronouns. Reliability studies for the annotation of the mentioned phenomena and for annotation of anaphoric links in general offer satisfactory results. The Wikipedia and blogs dataset will be distributed under Creative Commons Attributions licence.

We present a flexible toolkit-based approach to automatic coreference resolution on German text. We start with our previous work aimed at reimplementing the system from Soon et al. (2001) for English, and extend it to duplicate a version of the state-of-the-art proposal from Klenner and Ailloud (2009). Evaluation performed on a benchmarking dataset, namely the TueBa-D/Z corpus (Hinrichs et al., 2005b), shows that machine learning based coreference resolution can be robustly performed in a language other than English.

This paper summarizes our work on creating a full-scale coreference resolution (CR) system for Italian, using BART ― an open-source modular CR toolkit initially developed for English corpora. We discuss our experiments on language-specific issues of the task. As our evaluation experiments show, a language-agnostic system (designed primarily for English) can achieve a performance level in high forties (MUC F-score) when re-trained and tested on a new language, at least on gold mention boundaries. Compared to this level, we can improve our F-score by around 10% introducing a small number of language-specific changes. This shows that, with a modular coreference resolution platform, such as BART, one can straightforwardly develop a family of robust and reliable systems for various languages. We hope that our experiments will encourage researchers working on coreference in other languages to create their own full-scale coreference resolution systems ― as we have mentioned above, at the moment such modules exist only for very few languages other than English.

Recent years have seen an increasing interest in developing standards for linguistic annotation, with a focus on the interoperability of the resources. This effort, however, requires a profound knowledge of the advantages and disadvantages of linguistic annotation schemes in order to avoid importing the flaws and weaknesses of existing encoding schemes into the new standards. This paper addresses the question how to compare syntactically annotated corpora and gain insights into the usefulness of specific design decisions. We present an exhaustive evaluation of two German treebanks with crucially different encoding schemes. We evaluate three different parsers trained on the two treebanks and compare results using EvalB, the Leaf-Ancestor metric, and a dependency-based evaluation. Furthermore, we present TePaCoC, a new testsuite for the evaluation of parsers on complex German grammatical constructions. The testsuite provides a well thought-out error classification, which enables us to compare parser output for parsers trained on treebanks with different encoding schemes and provides interesting insights into the impact of treebank annotation schemes on specific constructions like PP attachment or non-constituent coordination.

Developing a full coreference system able to run all the way from raw text to semantic interpretation is a considerable engineering effort. Accordingly, there is very limited availability of off-the shelf tools for researchers whose interests are not primarily in coreference or others who want to concentrate on a specific aspect of the problem. We present BART, a highly modular toolkit for developing coreference applications. In the Johns Hopkins workshop on using lexical and encyclopedic knowledge for entity disambiguation, the toolkit was used to extend a reimplementation of Soon et al.’s proposal with a variety of additional syntactic and knowledge-based features, and experiment with alternative resolution processes, preprocessing tools, and classifiers. BART has been released as open source software and is available from http://www.sfs.uni-tuebingen.de/~versley/BART