The prediction of Environmental, Social, and Governance (ESG) impact and duration (length) of impact from company events, as reported in news articles, hold immense significance for investors, policymakers, and various stakeholders. In this paper, we describe solutions from our team “Upaya” to ESG impact and length prediction tasks on one such dataset ML-ESG-3. ML-ESG-3 dataset was released along with shared task as a part of the Fifth Workshop on Knowledge Discovery from Unstructured Data in Financial Services, co-located with LREC-COLING 2024. We employed two different paradigms to adapt Large Language Models (LLMs) to predict both the ESG impact and length of events. In the first approach, we leverage GPT-4 within the In-context learning (ICL) framework. A learning-free dense retriever identifies top K-relevant In-context learning examples from the training data for a given test example. The second approach involves instruction-tuning Mistral (7B) LLM to predict impact and duration, supplemented with rationale generated using GPT-4. Our models secured second place in French tasks and achieved reasonable results (fifth and ninth rank) in English tasks. These results demonstrate the potential of different LLM-based paradigms for delivering valuable insights within the ESG investing landscape.
Disambiguating a word’s intended meaning(sense) in a given context is important in Nat-ural Language Understanding (NLU). WSDaims to determine the correct sense of ambigu-ous words in context. At the same time, LMD(a WSD variation) focuses on disambiguatinglocation mention. Both tasks are vital in Nat-ural Language Processing (NLP) and informa-tion retrieval, as they help correctly interpretand extract information from text. Arabic ver-sion is further challenging because of its mor-phological richness, encompassing a complexinterplay of roots, stems, and affixes. This pa-per describes our solutions to both tasks, em-ploying Llama3 and Cohere-based models un-der Zero-Shot Learning and Re-Ranking, re-spectively. Both the shared tasks were partof the second Arabic Natural Language Pro-cessing Conference co-located with ACL 2024.Overall, we achieved 1st rank in the WSD task(accuracy 78%) and 2nd rank in the LMD task(MRR@1 0.59)
Reliable automatic evaluation of summarization systems is challenging due to the multifaceted and subjective nature of the task. This is especially the case for languages other than English, where human evaluations are scarce. In this work, we introduce SEAHORSE, a dataset for multilingual, multifaceted summarization evaluation. SEAHORSE consists of 96K summaries with human ratings along 6 dimensions of text quality: comprehensibility, repetition, grammar, attribution, main ideas, and conciseness, covering 6 languages, 9 systems, and 4 datasets. As a result of its size and scope, SEAHORSE can serve both as a benchmark to evaluate learnt metrics, as well as a large-scale resource for training such metrics. We show that metrics trained with SEAHORSE achieve strong performance on the out-of-domain meta-evaluation benchmarks TRUE (Honovich et al., 2022) and mFACE (Aharoni et al., 2022). We make the SEAHORSE dataset and metrics publicly available for future research on multilingual and multifaceted summarization evaluation.
Existing data-to-text generation datasets are mostly limited to English. To address this lack of data, we create Table-to-Text in African languages (TaTA), the first large multilingual table-to-text dataset with a focus on African languages. We created TaTA by transcribing figures and accompanying text in bilingual reports by the Demographic and Health Surveys Program, followed by professional translation to make the dataset fully parallel. TaTA includes 8,700 examples in nine languages including four African languages (Hausa, Igbo, Swahili, and Yorùbá) and a zero-shot test language (Russian). We additionally release screenshots of the original figures for future research on multilingual multi-modal approaches. Through an in-depth human evaluation, we show that TaTA is challenging for current models and that less than half the outputs from an mT5-XXL-based model are understandable and attributable to the source data. Our results highlight a) the need for validating metrics; and b) the importance of domain-specific metrics.
Relation extraction (RE) has achieved remarkable progress with the help of pre-trained language models. However, existing RE models are usually incapable of handling two situations: implicit expressions and long-tail relation classes, caused by language complexity and data sparsity. Further, these approaches and models are largely inaccessible to users who don’t have direct access to large language models (LLMs) and/or infrastructure for supervised training or fine-tuning. Rule-based systems also struggle with implicit expressions. Apart from this, Real world financial documents such as various 10-X reports (including 10-K, 10-Q, etc.) of publicly traded companies pose another challenge to rule-based systems in terms of longer and complex sentences. In this paper, we introduce a simple approach that consults training relations at test time through a nearest-neighbor search over dense vectors of lexico-syntactic patterns and provides a simple yet effective means to tackle the above issues. We evaluate our approach on REFinD and show that our method achieves state-of-the-art performance. We further show that it can provide a good start for human in the loop setup when a small number of annotations are available and it is also beneficial when domain experts can provide high quality patterns. Our code is available at 1.
To align conditional text generation model outputs with desired behaviors, there has been an increasing focus on training the model using reinforcement learning (RL) with reward functions learned from human annotations. Under this framework, we identify three common cases where high rewards are incorrectly assigned to undesirable patterns: noise-induced spurious correlation, naturally occurring spurious correlation, and covariate shift. We show that even though learned metrics achieve high performance on the distribution of the data used to train the reward function, the undesirable patterns may be amplified during RL training of the text generation model. While there has been discussion about reward gaming in the RL or safety community, in this discussion piece, we would like to highlight reward gaming in the natural language generation (NLG) community using concrete conditional text generation examples and discuss potential fixes and areas for future work.
Relation extraction (RE) is a crucial task in natural language processing (NLP) that aims to identify and classify relationships between entities mentioned in text. In the financial domain, relation extraction plays a vital role in extracting valuable information from financial documents, such as news articles, earnings reports, and company filings. This paper describes our solution to relation extraction on one such dataset REFinD. The dataset was released along with shared task as a part of the Fourth Workshop on Knowledge Discovery from Unstructured Data in Financial Services, co-located with SIGIR 2023. In this paper, we employed OpenAI models under the framework of in-context learning (ICL). We utilized two retrieval strategies to find top K relevant in-context learning demonstrations / examples from training data for a given test example. The first retrieval mechanism, we employed, is a learning-free dense retriever and the other system is a learning-based retriever. We were able to achieve 3rd rank overall. Our best F1-score is 0.718.
Data-to-text generation focuses on generating fluent natural language responses from structured meaning representations (MRs). Such representations are compositional and it is costly to collect responses for all possible combinations of atomic meaning schemata, thereby necessitating few-shot generalization to novel MRs. In this work, we systematically study the compositional generalization of the state-of-the-art T5 models in few-shot data-to-text tasks. We show that T5 models fail to generalize to unseen MRs, and we propose a template-based input representation that considerably improves the model’s generalization capability. To further improve the model’s performance, we propose an approach based on self-training using fine-tuned BLEURT for pseudo-response selection. On the commonly-used SGD and Weather benchmarks, the proposed self-training approach improves tree accuracy by 46%+ and reduces the slot error rates by 73%+ over the strong T5 baselines in few-shot settings.
Natural Language Processing (NLP) has achieved great progress in the past decade on the basis of neural models, which often make use of large amounts of labeled data to achieve state-of-the-art performance. The dependence on labeled data prevents NLP models from being applied to low-resource settings and languages because of the time, money, and expertise that is often required to label massive amounts of textual data. Consequently, the ability to learn with limited labeled data is crucial for deploying neural systems to real-world NLP applications. Recently, numerous approaches have been explored to alleviate the need for labeled data in NLP such as data augmentation and semi-supervised learning. This tutorial aims to provide a systematic and up-to-date overview of these methods in order to help researchers and practitioners understand the landscape of approaches and the challenges associated with learning from limited labeled data, an emerging topic in the computational linguistics community. We will consider applications to a wide variety of NLP tasks (including text classification, generation, and structured prediction) and will highlight current challenges and future directions.
We introduce GEM, a living benchmark for natural language Generation (NLG), its Evaluation, and Metrics. Measuring progress in NLG relies on a constantly evolving ecosystem of automated metrics, datasets, and human evaluation standards. Due to this moving target, new models often still evaluate on divergent anglo-centric corpora with well-established, but flawed, metrics. This disconnect makes it challenging to identify the limitations of current models and opportunities for progress. Addressing this limitation, GEM provides an environment in which models can easily be applied to a wide set of tasks and in which evaluation strategies can be tested. Regular updates to the benchmark will help NLG research become more multilingual and evolve the challenge alongside models. This paper serves as the description of the data for the 2021 shared task at the associated GEM Workshop.
Unsupervised translation has reached impressive performance on resource-rich language pairs such as English-French and English-German. However, early studies have shown that in more realistic settings involving low-resource, rare languages, unsupervised translation performs poorly, achieving less than 3.0 BLEU. In this work, we show that multilinguality is critical to making unsupervised systems practical for low-resource settings. In particular, we present a single model for 5 low-resource languages (Gujarati, Kazakh, Nepali, Sinhala, and Turkish) to and from English directions, which leverages monolingual and auxiliary parallel data from other high-resource language pairs via a three-stage training scheme. We outperform all current state-of-the-art unsupervised baselines for these languages, achieving gains of up to 14.4 BLEU. Additionally, we outperform strong supervised baselines for various language pairs as well as match the performance of the current state-of-the-art supervised model for Nepali-English. We conduct a series of ablation studies to establish the robustness of our model under different degrees of data quality, as well as to analyze the factors which led to the superior performance of the proposed approach over traditional unsupervised models.
We propose a straightforward vocabulary adaptation scheme to extend the language capacity of multilingual machine translation models, paving the way towards efficient continual learning for multilingual machine translation. Our approach is suitable for large-scale datasets, applies to distant languages with unseen scripts, incurs only minor degradation on the translation performance for the original language pairs and provides competitive performance even in the case where we only possess monolingual data for the new languages.
Recent developments in machine translation and multilingual text generation have led researchers to adopt trained metrics such as COMET or BLEURT, which treat evaluation as a regression problem and use representations from multilingual pre-trained models such as XLM-RoBERTa or mBERT. Yet studies on related tasks suggest that these models are most efficient when they are large, which is costly and impractical for evaluation. We investigate the trade-off between multilinguality and model capacity with RemBERT, a state-of-the-art multilingual language model, using data from the WMT Metrics Shared Task. We present a series of experiments which show that model size is indeed a bottleneck for cross-lingual transfer, then demonstrate how distillation can help addressing this bottleneck, by leveraging synthetic data generation and transferring knowledge from one teacher to multiple students trained on related languages. Our method yields up to 10.5% improvement over vanilla fine-tuning and reaches 92.6% of RemBERT’s performance using only a third of its parameters.
The quality of machine translation systems has dramatically improved over the last decade, and as a result, evaluation has become an increasingly challenging problem. This paper describes our contribution to the WMT 2020 Metrics Shared Task, the main benchmark for automatic evaluation of translation. We make several submissions based on BLEURT, a previously published which uses transfer learning. We extend the metric beyond English and evaluate it on 14 language pairs for which fine-tuning data is available, as well as 4 “zero-shot” language pairs, for which we have no labelled examples. Additionally, we focus on English to German and demonstrate how to combine BLEURT’s predictions with those of YiSi and use alternative reference translations to enhance the performance. Empirical results show that the models achieve competitive results on the WMT Metrics 2019 Shared Task, indicating their promise for the 2020 edition.
Text generation has made significant advances in the last few years. Yet, evaluation metrics have lagged behind, as the most popular choices (e.g., BLEU and ROUGE) may correlate poorly with human judgment. We propose BLEURT, a learned evaluation metric for English based on BERT. BLEURT can model human judgment with a few thousand possibly biased training examples. A key aspect of our approach is a novel pre-training scheme that uses millions of synthetic examples to help the model generalize. BLEURT provides state-of-the-art results on the last three years of the WMT Metrics shared task and the WebNLG data set. In contrast to a vanilla BERT-based approach, it yields superior results even when the training data is scarce and out-of-distribution.
We present ToTTo, an open-domain English table-to-text dataset with over 120,000 training examples that proposes a controlled generation task: given a Wikipedia table and a set of highlighted table cells, produce a one-sentence description. To obtain generated targets that are natural but also faithful to the source table, we introduce a dataset construction process where annotators directly revise existing candidate sentences from Wikipedia. We present systematic analyses of our dataset and annotation process as well as results achieved by several state-of-the-art baselines. While usually fluent, existing methods often hallucinate phrases that are not supported by the table, suggesting that this dataset can serve as a useful research benchmark for high-precision conditional text generation.
We present a probabilistic framework for multilingual neural machine translation that encompasses supervised and unsupervised setups, focusing on unsupervised translation. In addition to studying the vanilla case where there is only monolingual data available, we propose a novel setup where one language in the (source, target) pair is not associated with any parallel data, but there may exist auxiliary parallel data that contains the other. This auxiliary data can naturally be utilized in our probabilistic framework via a novel cross-translation loss term. Empirically, we show that our approach results in higher BLEU scores over state-of-the-art unsupervised models on the WMT’14 English-French, WMT’16 English-German, and WMT’16 English-Romanian datasets in most directions.
We present the Natural Questions corpus, a question answering data set. Questions consist of real anonymized, aggregated queries issued to the Google search engine. An annotator is presented with a question along with a Wikipedia page from the top 5 search results, and annotates a long answer (typically a paragraph) and a short answer (one or more entities) if present on the page, or marks null if no long/short answer is present. The public release consists of 307,373 training examples with single annotations; 7,830 examples with 5-way annotations for development data; and a further 7,842 examples with 5-way annotated sequestered as test data. We present experiments validating quality of the data. We also describe analysis of 25-way annotations on 302 examples, giving insights into human variability on the annotation task. We introduce robust metrics for the purposes of evaluating question answering systems; demonstrate high human upper bounds on these metrics; and establish baseline results using competitive methods drawn from related literature.
Existing open-domain question answering (QA) models are not suitable for real-time usage because they need to process several long documents on-demand for every input query, which is computationally prohibitive. In this paper, we introduce query-agnostic indexable representations of document phrases that can drastically speed up open-domain QA. In particular, our dense-sparse phrase encoding effectively captures syntactic, semantic, and lexical information of the phrases and eliminates the pipeline filtering of context documents. Leveraging strategies for optimizing training and inference time, our model can be trained and deployed even in a single 4-GPU server. Moreover, by representing phrases as pointers to their start and end tokens, our model indexes phrases in the entire English Wikipedia (up to 60 billion phrases) using under 2TB. Our experiments on SQuAD-Open show that our model is on par with or more accurate than previous models with 6000x reduced computational cost, which translates into at least 68x faster end-to-end inference benchmark on CPUs. Code and demo are available at nlp.cs.washington.edu/denspi
Automatically constructed datasets for generating text from semi-structured data (tables), such as WikiBio, often contain reference texts that diverge from the information in the corresponding semi-structured data. We show that metrics which rely solely on the reference texts, such as BLEU and ROUGE, show poor correlation with human judgments when those references diverge. We propose a new metric, PARENT, which aligns n-grams from the reference and generated texts to the semi-structured data before computing their precision and recall. Through a large scale human evaluation study of table-to-text models for WikiBio, we show that PARENT correlates with human judgments better than existing text generation metrics. We also adapt and evaluate the information extraction based evaluation proposed by Wiseman et al (2017), and show that PARENT has comparable correlation to it, while being easier to use. We show that PARENT is also applicable when the reference texts are elicited from humans using the data from the WebNLG challenge.
Generalization and reliability of multilingual translation often highly depend on the amount of available parallel data for each language pair of interest. In this paper, we focus on zero-shot generalization—a challenging setup that tests models on translation directions they have not been optimized for at training time. To solve the problem, we (i) reformulate multilingual translation as probabilistic inference, (ii) define the notion of zero-shot consistency and show why standard training often results in models unsuitable for zero-shot tasks, and (iii) introduce a consistent agreement-based training method that encourages the model to produce equivalent translations of parallel sentences in auxiliary languages. We test our multilingual NMT models on multiple public zero-shot translation benchmarks (IWSLT17, UN corpus, Europarl) and show that agreement-based learning often results in 2-3 BLEU zero-shot improvement over strong baselines without any loss in performance on supervised translation directions.
We propose a novel conditioned text generation model. It draws inspiration from traditional template-based text generation techniques, where the source provides the content (i.e., what to say), and the template influences how to say it. Building on the successful encoder-decoder paradigm, it first encodes the content representation from the given input text; to produce the output, it retrieves exemplar text from the training data as “soft templates,” which are then used to construct an exemplar-specific decoder. We evaluate the proposed model on abstractive text summarization and data-to-text generation. Empirical results show that this model achieves strong performance and outperforms comparable baselines.
We formalize a new modular variant of current question answering tasks by enforcing complete independence of the document encoder from the question encoder. This formulation addresses a key challenge in machine comprehension by building a standalone representation of the document discourse. It additionally leads to a significant scalability advantage since the encoding of the answer candidate phrases in the document can be pre-computed and indexed offline for efficient retrieval. We experiment with baseline models for the new task, which achieve a reasonable accuracy but significantly underperform unconstrained QA models. We invite the QA research community to engage in Phrase-Indexed Question Answering (PIQA, pika) for closing the gap. The leaderboard is at: nlp.cs.washington.edu/piqa