Use of synthetic data is rapidly emerging as a realistic alternative to manually annotating live traffic for industry-scale model building.
Here we propose an automatic red teaming framework that evaluates a given model and exposes its vulnerabilities against unsafe and inappropriate content generation.
In order to facilitate the research in multimodal FL, we introduce FedMultimodal, the first FL benchmark for multimodal learning covering five representative multimodal applications from ten commonly used datasets with a total of eight unique modalities.
We present a novel approach which uses prompt-tuning to control the extraction rates of memorized content in LLMs.
Transgender and non-binary (TGNB) individuals disproportionately experience discrimination and exclusion from daily life.
The estimated voltage sensitivity coefficients are used to model the nodal voltages, and the control robustness is achieved by accounting for their uncertainties.
As a use case, we leverage multilingual articles from two different data sources and build a first-of-its-kind multi-sentential code-mixed Hinglish dataset i. e., MUTANT.
First, we use this system to stress tests question answering, machine translation, and semantic parsing.
Natural language often contains ambiguities that can lead to misinterpretation and miscommunication.
Hypothesis rejection modules in both schemes reject/accept a hypothesis based on features drawn from the utterance directed to the SLU system, the associated SLU hypothesis and SLU confidence score.
We present a systematic analysis of the impact of decoding algorithms on LM fairness, and analyze the trade-off between fairness, diversity and quality.
1 code implementation • 2 Aug 2022 • Saleh Soltan, Shankar Ananthakrishnan, Jack FitzGerald, Rahul Gupta, Wael Hamza, Haidar Khan, Charith Peris, Stephen Rawls, Andy Rosenbaum, Anna Rumshisky, Chandana Satya Prakash, Mukund Sridhar, Fabian Triefenbach, Apurv Verma, Gokhan Tur, Prem Natarajan
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
Recent large-scale natural language processing (NLP) systems use a pre-trained Large Language Model (LLM) on massive and diverse corpora as a headstart.
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.
Training mixed-domain translation models is a complex task that demands tailored architectures and costly data preparation techniques.
Multiple metrics have been introduced to measure fairness in various natural language processing tasks.
Language models excel at generating coherent text, and model compression techniques such as knowledge distillation have enabled their use in resource-constrained settings.
With the rapid growth in language processing applications, fairness has emerged as an important consideration in data-driven solutions.
In this study, we evaluate the impact of such idiosyncrasies on Natural Language Understanding (NLU) models trained using FL.
This formulation is applied to control distributed controllable photovoltaic (PV) generation in a distribution network to restrict the voltage within prescribed limits.
Specifically, the proposed framework optimizes the dispatch plan of an upstream medium voltage (MV) grid accounting for the flexibility offered by downstream low voltage (LV) grids and the knowledge of the uncertainties of the stochastic resources.
There is an increasing interest in continuous learning (CL), as data privacy is becoming a priority for real-world machine learning applications.
This paper proposes and experimentally validates a joint control and scheduling framework for a grid-forming converter-interfaced BESS providing multiple services to the electrical grid.
However, the data used to train NLU models may contain private information such as addresses or phone numbers, particularly when drawn from human subjects.
Existing bias mitigation methods to reduce disparities in model outcomes across cohorts have focused on data augmentation, debiasing model embeddings, or adding fairness-based optimization objectives during training.
Increasing concerns and regulations about data privacy and sparsity necessitate the study of privacy-preserving, decentralized learning methods for natural language processing (NLP) tasks.
We prove the theoretical privacy guarantee of our algorithm and assess its privacy leakage under Membership Inference Attacks(MIA) (Shokri et al., 2017) on models trained with transformed data.
We make use of a conditional generator for data augmentation that is trained directly using the meta-learning objective and simultaneously with prototypical networks, hence ensuring that data augmentation is customized to the task.
To systematically study and benchmark social biases in open-ended language generation, we introduce the Bias in Open-Ended Language Generation Dataset (BOLD), a large-scale dataset that consists of 23, 679 English text generation prompts for bias benchmarking across five domains: profession, gender, race, religion, and political ideology.
To achieve a large terahertz (THz) amplitude from a spintronic THz emitter (STE), materials with 100\% spin polarisation such as Co-based Heusler compounds as the ferromagnetic layer are required.
Materials Science Mesoscale and Nanoscale Physics Other Condensed Matter Optics
Neural Architecture Search (NAS) methods, which automatically learn entire neural model or individual neural cell architectures, have recently achieved competitive or state-of-the-art (SOTA) performance on variety of natural language processing and computer vision tasks, including language modeling, natural language inference, and image classification.
We propose a novel framework, ADVIN, to automatically discover novel domains and intents from large volumes of unlabeled data.
In this research work, we aim to achieve classification parity across explicit as well as implicit sensitive features.
In this work we address this by proposing a generative model for multi-dimensional annotation fusion, which models the dimensions jointly leading to more accurate ground truth estimates.
To address the latency and computational complexity issues, we explore a BranchyNet scheme on an intent classification scheme within SLU systems.
In this work, we present NeuralBugLocator, a deep learning based technique, that can localize the bugs in a faulty program with respect to a failing test, without even running the program.
In this work, we experiment with variants of GAN architectures to generate feature vectors corresponding to an emotion in two ways: (i) A generator is trained with samples from a mixture prior.
An OVA system consists of as many OVA models as the number of classes, providing the advantage of asynchrony, where each OVA model can be re-trained independent of other models.
To localize the bugs, we analyze the trained network using a state-of-the-art neural prediction attribution technique and see which lines of the programs make it predict the test outcomes.
Convolutional Neural Networks (CNNs) have revolutionized performances in several machine learning tasks such as image classification, object tracking, and keyword spotting.
Sentiment analysis is a task that may suffer from a lack of data in certain cases, as the datasets are often generated and annotated by humans.
An ideal re-ranker will exhibit the following two properties: (a) it should prefer the most relevant hypothesis for the given input as the top hypothesis and, (b) the interpretation scores corresponding to each hypothesis produced by the re-ranker should be calibrated.
GANs consist of a discriminator and a generator working in tandem playing a min-max game to learn a target underlying data distribution; when fed with data-points sampled from a simpler distribution (like uniform or Gaussian distribution).
Sentiment classification involves quantifying the affective reaction of a human to a document, media item or an event.
Recently, generative adversarial networks and adversarial autoencoders have gained a lot of attention in machine learning community due to their exceptional performance in tasks such as digit classification and face recognition.
Novice programmers often struggle with the formal syntax of programming languages.
Ranked #4 on Program Repair on DeepFix
In this work, we propose Expectation-Maximization (EM) based algorithms that rely on the judgments from multiple annotators and the object attributes for inferring the latent ground truth.