Search Results for author:
Found 54 papers, 9 papers with code
Controlled Data Generation via Insertion Operations for NLU
Use of synthetic data is rapidly emerging as a realistic alternative to manually annotating live traffic for industry-scale model building.
Scalable and Accurate Self-supervised Multimodal Representation Learning without Aligned Video and Text Data
Currently popular video-text data mining approach via automatic speech recognition (ASR) used in HowTo100M provides low-quality captions that often do not refer to the video content.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+5
Larger Probes Tell a Different Story: Extending Psycholinguistic Datasets Via In-Context Learning
Language model probing is often used to test specific capabilities of these models.
Scaling Down to Scale Up: A Guide to Parameter-Efficient Fine-Tuning
This paper presents a systematic overview and comparison of parameter-efficient fine-tuning methods covering over 40 papers published between February 2019 and February 2023.
Reasoning Circuits: Few-shot Multihop Question Generation with Structured Rationales
Multi-hop Question Generation is the task of generating questions which require the reader to reason over and combine information spread across multiple passages using several reasoning steps.
On Task-Adaptive Pretraining for Dialogue Response Selection
Recent advancements in dialogue response selection (DRS) are based on the \textit{task-adaptive pre-training (TAP)} approach, by first initializing their model with BERT~\cite{devlin-etal-2019-bert}, and adapt to dialogue data with dialogue-specific or fine-grained pre-training tasks.
AlexaTM 20B: Few-Shot Learning Using a Large-Scale Multilingual Seq2Seq Model
In this work, we demonstrate that multilingual large-scale sequence-to-sequence (seq2seq) models, pre-trained on a mixture of denoising and Causal Language Modeling (CLM) tasks, are more efficient few-shot learners than decoder-only models on various tasks.
Ranked #8 on
Natural Language Inference
on CommitmentBank
Learning to Ask Like a Physician
The questions are generated by medical experts from 100+ MIMIC-III discharge summaries.
Life after BERT: What do Other Muppets Understand about Language?
Existing pre-trained transformer analysis works usually focus only on one or two model families at a time, overlooking the variability of the architecture and pre-training objectives.
Down and Across: Introducing Crossword-Solving as a New NLP Benchmark
Solving crossword puzzles requires diverse reasoning capabilities, access to a vast amount of knowledge about language and the world, and the ability to satisfy the constraints imposed by the structure of the puzzle.
Natural Language Understanding
Open-Domain Question Answering
+1
Federated Learning with Noisy User Feedback
Federated learning (FL) has recently emerged as a method for training ML models on edge devices using sensitive user data and is seen as a way to mitigate concerns over data privacy.
Multi-Stream Transformers
Transformer-based encoder-decoder models produce a fused token-wise representation after every encoder layer.
An Efficient DP-SGD Mechanism for Large Scale NLP Models
However, the data used to train NLU models may contain private information such as addresses or phone numbers, particularly when drawn from human subjects.
BERT Busters: Outlier Dimensions that Disrupt Transformers
Multiple studies have shown that Transformers are remarkably robust to pruning.
A guide to the dataset explosion in QA, NLI, and commonsense reasoning
Question answering, natural language inference and commonsense reasoning are increasingly popular as general NLP system benchmarks, driving both modeling and dataset work.
Update Frequently, Update Fast: Retraining Semantic Parsing Systems in a Fraction of Time
To reduce training time, one can fine-tune the previously trained model on each patch, but naive fine-tuning exhibits catastrophic forgetting - degradation of the model performance on the data not represented in the data patch.
Towards Visual Dialog for Radiology
Using MIMIC-CXR, an openly available database of chest X-ray images, we construct both a synthetic and a real-world dataset and provide baseline scores achieved by state-of-the-art models.
When BERT Plays the Lottery, All Tickets Are Winning
Large Transformer-based models were shown to be reducible to a smaller number of self-attention heads and layers.
A Primer in BERTology: What we know about how BERT works
Transformer-based models have pushed state of the art in many areas of NLP, but our understanding of what is behind their success is still limited.
Calls to Action on Social Media: Detection, Social Impact, and Censorship Potential
Calls to action on social media are known to be effective means of mobilization in social movements, and a frequent target of censorship.
Memory-Augmented Recurrent Networks for Dialogue Coherence
The first design features a sequence-to-sequence architecture with two separate NTM modules, one for each participant in the conversation.
Injecting Hierarchy with U-Net Transformers
The Transformer architecture has become increasingly popular over the past two years, owing to its impressive performance on a number of natural language processing (NLP) tasks.
NarrativeTime: Dense Temporal Annotation on a Timeline
For the past decade, temporal annotation has been sparse: only a small portion of event pairs in a text was annotated.
Solving Math Word Problems with Double-Decoder Transformer
This paper proposes a Transformer-based model to generate equations for math word problems.
Revealing the Dark Secrets of BERT
BERT-based architectures currently give state-of-the-art performance on many NLP tasks, but little is known about the exact mechanisms that contribute to its success.
What's in a Name? Reducing Bias in Bios without Access to Protected Attributes
In the context of mitigating bias in occupation classification, we propose a method for discouraging correlation between the predicted probability of an individual's true occupation and a word embedding of their name.
Adversarial Text Generation Without Reinforcement Learning
This is largely because sequences of text are discrete, and thus gradients cannot propagate from the discriminator to the generator.
Similarity-Based Reconstruction Loss for Meaning Representation
This paper addresses the problem of representation learning.
Triad-based Neural Network for Coreference Resolution
We propose a triad-based neural network system that generates affinity scores between entity mentions for coreference resolution.
Adversarial Decomposition of Text Representation
We show that the proposed method is capable of fine-grained controlled change of these aspects of the input sentence.
What's in Your Embedding, And How It Predicts Task Performance
Attempts to find a single technique for general-purpose intrinsic evaluation of word embeddings have so far not been successful.
RuSentiment: An Enriched Sentiment Analysis Dataset for Social Media in Russian
This paper presents RuSentiment, a new dataset for sentiment analysis of social media posts in Russian, and a new set of comprehensive annotation guidelines that are extensible to other languages.
Ranked #2 on
Sentiment Analysis
on RuSentiment
Context-Aware Neural Model for Temporal Information Extraction
We propose a context-aware neural network model for temporal information extraction.
CliNER 2.0: Accessible and Accurate Clinical Concept Extraction
Clinical notes often describe important aspects of a patient's stay and are therefore critical to medical research.
Forced Apart: Discovering Disentangled Representations Without Exhaustive Labels
Learning a better representation with neural networks is a challenging problem, which has been tackled from different perspectives in the past few years.
Length, Interchangeability, and External Knowledge: Observations from Predicting Argument Convincingness
In this work, we provide insight into three key aspects related to predicting argument convincingness.
Tracking Bias in News Sources Using Social Media: the Russia-Ukraine Maidan Crisis of 2013--2014
We show that on the task of predicting which side is likely to prefer a given article, a Naive Bayes classifier can record 90. 3{\%} accuracy looking only at domain names of the news sources.
Towards Debate Automation: a Recurrent Model for Predicting Debate Winners
In this paper we introduce a practical first step towards the creation of an automated debate agent: a state-of-the-art recurrent predictive model for predicting debate winners.
SemEval-2017 Task 6: \#HashtagWars: Learning a Sense of Humor
This paper describes a new shared task for humor understanding that attempts to eschew the ubiquitous binary approach to humor detection and focus on comparative humor ranking instead.
HumorHawk at SemEval-2017 Task 6: Mixing Meaning and Sound for Humor Recognition
This paper describes the winning system for SemEval-2017 Task 6: {\#}HashtagWars: Learning a Sense of Humor.
Forced to Learn: Discovering Disentangled Representations Without Exhaustive Labels
Learning a better representation with neural networks is a challenging problem, which was tackled extensively from different prospectives in the past few years.
Temporal Information Extraction for Question Answering Using Syntactic Dependencies in an LSTM-based Architecture
In this paper, we propose to use a set of simple, uniform in architecture LSTM-based models to recover different kinds of temporal relations from text.
Here's My Point: Joint Pointer Architecture for Argument Mining
One of the major goals in automated argumentation mining is to uncover the argument structure present in argumentative text.
Evaluating Creative Language Generation: The Case of Rap Lyric Ghostwriting
The goal of this paper is to develop evaluation methods for one such task, ghostwriting of rap lyrics, and to provide an explicit, quantifiable foundation for the goals and future directions of this task.
#HashtagWars: Learning a Sense of Humor
Our best supervised system achieved 63. 7% accuracy, suggesting that this task is much more difficult than comparable humor detection tasks.
Normalization of Relative and Incomplete Temporal Expressions in Clinical Narratives
We analyze the RI-TIMEXes in temporally annotated corpora and propose two hypotheses regarding the normalization of RI-TIMEXes in the clinical narrative domain: the anchor point hypothesis and the anchor relation hypothesis.
Word Sense Inventories by Non-Experts.
In this paper, we explore different strategies for implementing a crowdsourcing methodology for a single-step construction of an empirically-derived sense inventory and the corresponding sense-annotated corpus.
