Search Results for author:
Found 82 papers, 44 papers with code
Findings of the WMT 2020 Shared Task on Machine Translation Robustness
We report the findings of the second edition of the shared task on improving robustness in Machine Translation (MT).
Understanding Arithmetic Reasoning in Language Models using Causal Mediation Analysis
Our findings provide insights into the mechanistic interpretation of LLMs for arithmetic tasks and highlight the specific components involved in arithmetic reasoning.
Interpreting Transformer's Attention Dynamic Memory and Visualizing the Semantic Information Flow of GPT
In this paper, we interpret LM attention heads and memory values, the vectors the models dynamically create and recall while processing a given input.
Shielded Representations: Protecting Sensitive Attributes Through Iterative Gradient-Based Projection
In this work, we propose Iterative Gradient-Based Projection (IGBP), a novel method for removing non-linear encoded concepts from neural representations.
ContraSim -- A Similarity Measure Based on Contrastive Learning
In contrast to common closed-form similarity measures, ContraSim learns a parameterized measure by using both similar and dissimilar examples.
Editing Implicit Assumptions in Text-to-Image Diffusion Models
Our Text-to-Image Model Editing method, TIME for short, receives a pair of inputs: a "source" under-specified prompt for which the model makes an implicit assumption (e. g., "a pack of roses"), and a "destination" prompt that describes the same setting, but with a specified desired attribute (e. g., "a pack of blue roses").
Parallel Context Windows for Large Language Models
The key to the approach is to carve a long context into chunks (``windows''), restrict the attention mechanism to apply only within each window, and re-use the positional embeddings across the windows.
Debiasing NLP Models Without Demographic Information
Models trained from real-world data tend to imitate and amplify social biases.
What Are You Token About? Dense Retrieval as Distributions Over the Vocabulary
We leverage this insight and propose a simple way to enrich query and passage representations with lexical information at inference time, and show that this significantly improves performance compared to the original model in zero-shot settings, and specifically on the BEIR benchmark.
BLOOM: A 176B-Parameter Open-Access Multilingual Language Model
Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions.
Emergent Quantized Communication
The field of emergent communication aims to understand the characteristics of communication as it emerges from artificial agents solving tasks that require information exchange.
Choose Your Lenses: Flaws in Gender Bias Evaluation
In this position paper, we assess the current paradigm of gender bias evaluation and identify several flaws in it.
Measures of Information Reflect Memorization Patterns
In this work, we hypothesize -- and subsequently show -- that the diversity in the activation patterns of different neurons is reflective of model generalization and memorization.
Mass-Editing Memory in a Transformer
Recent work has shown exciting promise in updating large language models with new memories, so as to replace obsolete information or add specialized knowledge.
Measuring Causal Effects of Data Statistics on Language Model's `Factual' Predictions
Our causal framework and our results demonstrate the importance of studying datasets and the benefits of causality for understanding NLP models.
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models.
IDANI: Inference-time Domain Adaptation via Neuron-level Interventions
We propose a new approach for domain adaptation (DA), using neuron-level interventions: We modify the representation of each test example in specific neurons, resulting in a counterfactual example from the source domain, which the model is more familiar with.
MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning
Huge language models (LMs) have ushered in a new era for AI, serving as a gateway to natural-language-based knowledge tasks.
How Gender Debiasing Affects Internal Model Representations, and Why It Matters
Common studies of gender bias in NLP focus either on extrinsic bias measured by model performance on a downstream task or on intrinsic bias found in models' internal representations.
A Multilingual Perspective Towards the Evaluation of Attribution Methods in Natural Language Inference
Most evaluations of attribution methods focus on the English language.
Locating and Editing Factual Associations in GPT
To test our hypothesis that these computations correspond to factual association recall, we modify feed-forward weights to update specific factual associations using Rank-One Model Editing (ROME).
IRM---when it works and when it doesn't: A test case of natural language inference
Invariant Risk Minimization (IRM) is a recently proposed framework for out-of-distribution (o. o. d) generalization.
On the Pitfalls of Analyzing Individual Neurons in Language Models
Among these, the common approach is to use an external probe to rank neurons according to their relevance to some linguistic attribute, and to evaluate the obtained ranking using the same probe that produced it.
Variance Pruning: Pruning Language Models via Temporal Neuron Variance
As language models become larger, different pruning methods have been proposed to reduce model size.
Debiasing Methods in Natural Language Understanding Make Bias More Accessible
Model robustness to bias is often determined by the generalization on carefully designed out-of-distribution datasets.
A Generative Approach for Mitigating Structural Biases in Natural Language Inference
Many natural language inference (NLI) datasets contain biases that allow models to perform well by only using a biased subset of the input, without considering the remainder features.
Causal Analysis of Syntactic Agreement Mechanisms in Neural Language Models
Targeted syntactic evaluations have demonstrated the ability of language models to perform subject-verb agreement given difficult contexts.
Variational Information Bottleneck for Effective Low-Resource Fine-Tuning
Moreover, we show that our VIB model finds sentence representations that are more robust to biases in natural language inference datasets, and thereby obtains better generalization to out-of-domain datasets.
Supervising Model Attention with Human Explanations for Robust Natural Language Inference
Natural Language Inference (NLI) models are known to learn from biases and artefacts within their training data, impacting how well they generalise to other unseen datasets.
Probing Classifiers: Promises, Shortcomings, and Advances
Probing classifiers have emerged as one of the prominent methodologies for interpreting and analyzing deep neural network models of natural language processing.
Learning from others' mistakes: Avoiding dataset biases without modeling them
State-of-the-art natural language processing (NLP) models often learn to model dataset biases and surface form correlations instead of features that target the intended underlying task.
Investigating Gender Bias in Language Models Using Causal Mediation Analysis
As a case study, we apply this methodology to analyzing gender bias in pre-trained Transformer language models.
Similarity Analysis of Self-Supervised Speech Representations
We also design probing tasks to study the correlation between the models' pre-training loss and the amount of specific speech information contained in their learned representations.
Analyzing Individual Neurons in Pre-trained Language Models
We found small subsets of neurons to predict linguistic tasks, with lower level tasks (such as morphology) localized in fewer neurons, compared to higher level task of predicting syntax.
Interpretability and Analysis in Neural NLP
While deep learning has transformed the natural language processing (NLP) field and impacted the larger computational linguistics community, the rise of neural networks is stained by their opaque nature: It is challenging to interpret the inner workings of neural network models, and explicate their behavior.
Probing Neural Dialog Models for Conversational Understanding
The predominant approach to open-domain dialog generation relies on end-to-end training of neural models on chat datasets.
The Sensitivity of Language Models and Humans to Winograd Schema Perturbations
Large-scale pretrained language models are the major driving force behind recent improvements in performance on the Winograd Schema Challenge, a widely employed test of common sense reasoning ability.
Similarity Analysis of Contextual Word Representation Models
We use existing and novel similarity measures that aim to gauge the level of localization of information in the deep models, and facilitate the investigation of which design factors affect model similarity, without requiring any external linguistic annotation.
Probing the Probing Paradigm: Does Probing Accuracy Entail Task Relevance?
Although neural models have achieved impressive results on several NLP benchmarks, little is understood about the mechanisms they use to perform language tasks.
Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias
Common methods for interpreting neural models in natural language processing typically examine either their structure or their behavior, but not both.
Analyzing Redundancy in Pretrained Transformer Models
Transformer-based deep NLP models are trained using hundreds of millions of parameters, limiting their applicability in computationally constrained environments.
Memory-Augmented Recurrent Neural Networks Can Learn Generalized Dyck Languages
We introduce three memory-augmented Recurrent Neural Networks (MARNNs) and explore their capabilities on a series of simple language modeling tasks whose solutions require stack-based mechanisms.
On the Linguistic Representational Power of Neural Machine Translation Models
(iii) Do the representations capture lexical semantics?
A Constructive Prediction of the Generalization Error Across Scales
In this work, we present a functional form which approximates well the generalization error in practice.
End-to-End Bias Mitigation by Modelling Biases in Corpora
We experiment on large-scale natural language inference and fact verification benchmarks, evaluating on out-of-domain datasets that are specifically designed to assess the robustness of models against known biases in the training data.
On Adversarial Removal of Hypothesis-only Bias in Natural Language Inference
Popular Natural Language Inference (NLI) datasets have been shown to be tainted by hypothesis-only biases.
Analyzing Phonetic and Graphemic Representations in End-to-End Automatic Speech Recognition
End-to-end neural network systems for automatic speech recognition (ASR) are trained from acoustic features to text transcriptions.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
Don't Take the Premise for Granted: Mitigating Artifacts in Natural Language Inference
In contrast to standard approaches to NLI, our methods predict the probability of a premise given a hypothesis and NLI label, discouraging models from ignoring the premise.
Findings of the First Shared Task on Machine Translation Robustness
We share the findings of the first shared task on improving robustness of Machine Translation (MT).
Adversarial Regularization for Visual Question Answering: Strengths, Shortcomings, and Side Effects
Visual question answering (VQA) models have been shown to over-rely on linguistic biases in VQA datasets, answering questions "blindly" without considering visual context.
LSTM Networks Can Perform Dynamic Counting
In this paper, we systematically assess the ability of standard recurrent networks to perform dynamic counting and to encode hierarchical representations.
Analyzing the Structure of Attention in a Transformer Language Model
The Transformer is a fully attention-based alternative to recurrent networks that has achieved state-of-the-art results across a range of NLP tasks.
Improving Neural Language Models by Segmenting, Attending, and Predicting the Future
In this work, we propose a method that improves language modeling by learning to align the given context and the following phrase.
Ranked #15 on
Language Modelling
on WikiText-103
One Size Does Not Fit All: Comparing NMT Representations of Different Granularities
Recent work has shown that contextualized word representations derived from neural machine translation are a viable alternative to such from simple word predictions tasks.
Linguistic Knowledge and Transferability of Contextual Representations
Contextual word representations derived from large-scale neural language models are successful across a diverse set of NLP tasks, suggesting that they encode useful and transferable features of language.
Character-based Surprisal as a Model of Reading Difficulty in the Presence of Error
Also, transpositions are more difficult than misspellings, and a high error rate increases difficulty for all words, including correct ones.
Analysis Methods in Neural Language Processing: A Survey
The field of natural language processing has seen impressive progress in recent years, with neural network models replacing many of the traditional systems.
What Is One Grain of Sand in the Desert? Analyzing Individual Neurons in Deep NLP Models
We further present a comprehensive analysis of neurons with the aim to address the following questions: i) how localized or distributed are different linguistic properties in the models?
NeuroX: A Toolkit for Analyzing Individual Neurons in Neural Networks
We present a toolkit to facilitate the interpretation and understanding of neural network models.
Identifying and Controlling Important Neurons in Neural Machine Translation
Neural machine translation (NMT) models learn representations containing substantial linguistic information.
On Evaluating the Generalization of LSTM Models in Formal Languages
Recurrent Neural Networks (RNNs) are theoretically Turing-complete and established themselves as a dominant model for language processing.
Studying the History of the Arabic Language: Language Technology and a Large-Scale Historical Corpus
Arabic is a widely-spoken language with a long and rich history, but existing corpora and language technology focus mostly on modern Arabic and its varieties.
On the Evaluation of Semantic Phenomena in Neural Machine Translation Using Natural Language Inference
We propose a process for investigating the extent to which sentence representations arising from neural machine translation (NMT) systems encode distinct semantic phenomena.
Evaluating Layers of Representation in Neural Machine Translation on Part-of-Speech and Semantic Tagging Tasks
In this paper, we investigate the representations learned at different layers of NMT encoders.
Synthetic and Natural Noise Both Break Neural Machine Translation
Character-based neural machine translation (NMT) models alleviate out-of-vocabulary issues, learn morphology, and move us closer to completely end-to-end translation systems.
Understanding and Improving Morphological Learning in the Neural Machine Translation Decoder
End-to-end training makes the neural machine translation (NMT) architecture simpler, yet elegant compared to traditional statistical machine translation (SMT).
Analyzing Hidden Representations in End-to-End Automatic Speech Recognition Systems
In this work, we analyze the speech representations learned by a deep end-to-end model that is based on convolutional and recurrent layers, and trained with a connectionist temporal classification (CTC) loss.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
Challenging Language-Dependent Segmentation for Arabic: An Application to Machine Translation and Part-of-Speech Tagging
Word segmentation plays a pivotal role in improving any Arabic NLP application.
Neural Machine Translation Training in a Multi-Domain Scenario
Model stacking works best when training begins with the furthest out-of-domain data and the model is incrementally fine-tuned with the next furthest domain and so on.
What do Neural Machine Translation Models Learn about Morphology?
Neural machine translation (MT) models obtain state-of-the-art performance while maintaining a simple, end-to-end architecture.
Shamela: A Large-Scale Historical Arabic Corpus
Arabic is a widely-spoken language with a rich and long history spanning more than fourteen centuries.
Neural Attention for Learning to Rank Questions in Community Question Answering
In real-world data, e. g., from Web forums, text is often contaminated with redundant or irrelevant content, which leads to introducing noise in machine learning algorithms.
Large-Scale Machine Translation between Arabic and Hebrew: Available Corpora and Initial Results
Machine translation between Arabic and Hebrew has so far been limited by a lack of parallel corpora, despite the political and cultural importance of this language pair.
A Character-level Convolutional Neural Network for Distinguishing Similar Languages and Dialects
Discriminating between closely-related language varieties is considered a challenging and important task.
Fine-grained Analysis of Sentence Embeddings Using Auxiliary Prediction Tasks
The analysis sheds light on the relative strengths of different sentence embedding methods with respect to these low level prediction tasks, and on the effect of the encoded vector's dimensionality on the resulting representations.
Exploring Compositional Architectures and Word Vector Representations for Prepositional Phrase Attachment
In this paper, we show that word vector representations can yield significant PP attachment performance gains.
