Skip to search formSkip to main contentSkip to account menu

Deep Speech 2 : End-to-End Speech Recognition in English and Mandarin

It is shown that an end-to-end deep learning approach can be used to recognize either English or Mandarin Chinese speech-two vastly different languages, and is competitive with the transcription of human workers when benchmarked on standard datasets.
View PDF on arXiv (opens in a new tab)

The most dangerous code in the world: validating SSL certificates in non-browser software

It is demonstrated that SSL certificate validation is completely broken in many security-critical applications and libraries and badly designed APIs of SSL implementations and data-transport libraries which present developers with a confusing array of settings and options are analyzed.
View on ACM (opens in a new tab)

Structured Prediction as Translation between Augmented Natural Languages

This work proposes a new framework, Translation between Augmented Natural Languages (TANL), to solve many structured prediction language tasks including joint entity and relation extraction, nested named entity recognition, relation classification, semantic role labeling, event extraction, coreference resolution, and dialogue state tracking.
View PDF on arXiv (opens in a new tab)

Label Semantics for Few Shot Named Entity Recognition

The semantic information in the names of the labels are leveraged as a way of giving the model additional signal and enriched priors and the label semantics signal is shown to support improved state-of-the-art results in multiple few shot NER benchmarks and on-par performance in standard benchmarks.
View on ACL (opens in a new tab)

Resource-Enhanced Neural Model for Event Argument Extraction

This work proposes a trigger-aware sequence encoder with several types of trigger-dependent sequence representations and shows that the approach achieves a new state-of-the-art on the English ACE 2005 benchmark.
View on ACL (opens in a new tab)

Severing the Edge between before and after: Neural Architectures for Temporal Ordering of Events

A neural architecture and a set of training methods for ordering events by predicting temporal relations are proposed and experiments on the MATRES dataset of English documents establish a new state-of-the-art on this task.
View on ACL (opens in a new tab)

To BERT or Not to BERT: Comparing Task-specific and Task-agnostic Semi-Supervised Approaches for Sequence Tagging

This work investigates how to effectively use unlabeled data by exploring the task-specific semi-supervised approach, Cross-View Training (CVT) and comparing it with task-agnostic BERT in multiple settings that include domain and task relevant English data.
View on ACL (opens in a new tab)

Multi-Task Learning and Adapted Knowledge Models for Emotion-Cause Extraction

This work proposes novel methods that combine common-sense knowledge via adapted knowledge models with multi-task learning to perform joint emotion classification and emotion cause tagging and shows performance improvement on both tasks when including common- sense reasoning and a multitask framework.
View on ACL (opens in a new tab)

Instruction Tuning for Few-Shot Aspect-Based Sentiment Analysis

This work fine-tune a T5 model with instructional prompts in a multi-task learning fashion covering all the sub-tasks, as well as the entire quadruple prediction task, and proposes a unified framework for solving ABSA, and the associated sub-Tasks to improve the performance in few-shot scenarios.
View PDF on arXiv (opens in a new tab)

1 Contextual-window based neural network model for predicting rainfall in rural India

We propose a neural-network based model for predicting rainfall and subsequently, the occurrence of floods for any given city in India. Our initial implementation has primary focus on a rural city in…
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)