Leveraging both labeled (input-output associations) and unlabeled data (wider contextual grounding) may provide complementary benefits in retrieval augmented generation (RAG). However, effectively combining evidence from these heterogeneous sources is challenging as the respective similarity scores are not inter-comparable. Additionally, aggregating beliefs from the outputs of multiple rankers can improve the effectiveness of RAG. Our proposed method first aggregates the top-documents from a number of IR models using a standard rank fusion technique for each source (labeled and unlabeled). Next, we standardize the retrieval score distributions within each source by applying z-score transformation before merging the top-retrieved documents from the two sources. We evaluate our approach on the fact verification task, demonstrating that it consistently improves over the best-performing individual ranker or source and also shows better out-of-domain generalization.

Applications of Natural Language Processing (NLP) in the finance domain have been very popular of late. For financial NLP, (FinNLP) while various datasets exist for widely spoken languages like English and Chinese, datasets are scarce for low resource languages,particularly for Indian languages. In this paper, we address this challenges by presenting IndicFinNLP – a collection of 9 datasets consisting of three tasks relating to FinNLP for three Indian languages. These tasks are Exaggerated Numeral Detection, Sustainability Classification, and ESG Theme Determination of financial texts in Hindi, Bengali, and Telugu. Moreover, we release the datasets under CC BY-NC-SA 4.0 license for the benefit of the research community.

Extraction of methodology component names from scientific articles is a challenging task due to the diversified contexts around the occurrences of these entities, and the different levels of granularity and containment relationships exhibited by these entities. We hypothesize that standard sequence labeling approaches may not adequately model the dependence of methodology name mentions with their contexts, due to the problems of their large, fast evolving, and domain-specific vocabulary. As a solution, we propose a factored approach, where the mention-context dependencies are represented in a more fine-grained manner, thus allowing the model parameters to better adjust to the different characteristic patterns inherent within the data. In particular, we experiment with two variants of this factored approach - one that uses the per-entity category information derived from an ontology, and the other that makes use of the topology of the sentence embedding space to infer a category for each entity constituting that sentence. We demonstrate that both these factored variants of SciBERT outperform their non-factored counterpart, a state-of-the-art model for scientific concept extraction.

In automatic post-editing (APE) it makes sense to condition post-editing (pe) decisions on both the source (src) and the machine translated text (mt) as input.This has led to multi-encoder based neural APE approaches.A research challenge now is the search for architectures that best support the capture, preparation and provision of src and mt information and its integration with pe decisions.In this paper we present an efficient multi-encoder based APE model, called transference.Unlike previous approaches, it (i) uses a transformer encoder block for src, (ii) followed by a decoder block, but without masking for self-attention on mt, which effectively acts as second encoder combining src → mt, and (iii) feeds this representation into a final decoder block generating pe.Our model outperforms the best performing systems by 1 BLEU point on the WMT 2016, 2017, and 2018 English-German APE shared tasks (PBSMT and NMT).Furthermore, the results of our model on the WMT 2019 APE task using NMT data shows performance at the level of the state-of-the-art.The inference time of our model is similar to the vanilla transformer-based NMT system although our model deals with two separate encoders.We further investigate the importance of our newly introduced second encoder and find that decreasing the number of layers hurts performance, while reducing the number of layers of the decoder does not matter much.

Most text-simplification systems require an indicator of the complexity of the words. The prevalent approaches to word difficulty prediction are based on manual feature engineering. Using deep learning based models are largely left unexplored due to their comparatively poor performance. In this paper we explore the use of one of such in predicting the difficulty of words. We treat the problem as a binary classification problem. We train traditional machine learning models and evaluate their performance on the task. Removing dependency on frequency of previously acquired words for measuring difficulty was one of our primary aims. Then we analyze a convolutional neural network based prediction model which operates at the character level and evaluate its efficiency compared to others.

The field of speech recognition has seen tremendous advances in the recent past owing to the development of powerful deep learning architectures. However, the closely related fields of speech segmentation and di-arization are still primarily dominated by sophisticated variants of hierarchical clustering algorithms. We propose a powerful adaptation of the state-of-the-art Speech Recognition models for these tasks and demonstrate the effectiveness of our techniques on standard datasets. Our architectures are a combination of Bidirectional Long Short Term Memory (LSTM) Networks, Convolutional Networks, and Fully Connected Networks, trained by Gradient Descent to minimize the Cross Entropy and the Connectionist Temporal Classification (CTC) losses. We adapt the Libri Speech corpus for the task of segmentation and diarization. We obtained comparable results with respect to state-of-the-art in both tasks.

In this paper we combine two strands of machine translation (MT) research: automatic post-editing (APE) and multi-engine (system combination) MT. APE systems learn a target-language-side second stage MT system from the data produced by human corrected output of a first stage MT system, to improve the output of the first stage MT in what is essentially a sequential MT system combination architecture. At the same time, there is a rich research literature on parallel MT system combination where the same input is fed to multiple engines and the best output is selected or smaller sections of the outputs are combined to obtain improved translation output. In the paper we show that parallel system combination in the APE stage of a sequential MT-APE combination yields substantial translation improvements both measured in terms of automatic evaluation metrics as well as in terms of productivity improvements measured in a post-editing experiment. We also show that system combination on the level of APE alignments yields further improvements. Overall our APE system yields statistically significant improvement of 5.9% relative BLEU over a strong baseline (English–Italian Google MT) and 21.76% productivity increase in a human post-editing experiment with professional translators.