Grusha Prasad

Bio: Grusha Prasad is an academic researcher from Johns Hopkins University. The author has contributed to research in topics: Language model & Sentence. The author has an hindex of 6, co-authored 11 publications receiving 94 citations.

Dynabench: Rethinking Benchmarking in NLP.

Abstract: We introduce Dynabench, an open-source platform for dynamic dataset creation and model benchmarking. Dynabench runs in a web browser and supports human-and-model-in-the-loop dataset creation: annotators seek to create examples that a target model will misclassify, but that another person will not. In this paper, we argue that Dynabench addresses a critical need in our community: contemporary models quickly achieve outstanding performance on benchmark tasks but nonetheless fail on simple challenge examples and falter in real-world scenarios. With Dynabench, dataset creation, model development, and model assessment can directly inform each other, leading to more robust and informative benchmarks. We report on four initial NLP tasks, illustrating these concepts and highlighting the promise of the platform, and address potential objections to dynamic benchmarking as a new standard for the field.

Using Priming to Uncover the Organization of Syntactic Representations in Neural Language Models

Abstract: Neural language models (LMs) perform well on tasks that require sensitivity to syntactic structure. Drawing on the syntactic priming paradigm from psycholinguistics, we propose a novel technique to analyze the representations that enable such success. By establishing a gradient similarity metric between structures, this technique allows us to reconstruct the organization of the LMs’ syntactic representational space. We use this technique to demonstrate that LSTM LMs’ representations of different types of sentences with relative clauses are organized hierarchically in a linguistically interpretable manner, suggesting that the LMs track abstract properties of the sentence.

Using Priming to Uncover the Organization of Syntactic Representations in Neural Language Models

Abstract: Neural language models (LMs) perform well on tasks that require sensitivity to syntactic structure. Drawing on the syntactic priming paradigm from psycholinguistics, we propose a novel technique to analyze the representations that enable such success. By establishing a gradient similarity metric between structures, this technique allows us to reconstruct the organization of the LMs' syntactic representational space. We use this technique to demonstrate that LSTM LMs' representations of different types of sentences with relative clauses are organized hierarchically in a linguistically interpretable manner, suggesting that the LMs track abstract properties of the sentence.

Dynabench: Rethinking Benchmarking in NLP

Abstract: We introduce Dynabench, an open-source platform for dynamic dataset creation and model benchmarking. Dynabench runs in a web browser and supports human-and-model-in-the-loop dataset creation: annotators seek to create examples that a target model will misclassify, but that another person will not. In this paper, we argue that Dynabench addresses a critical need in our community: contemporary models quickly achieve outstanding performance on benchmark tasks but nonetheless fail on simple challenge examples and falter in real-world scenarios. With Dynabench, dataset creation, model development, and model assessment can directly inform each other, leading to more robust and informative benchmarks. We report on four initial NLP tasks, illustrating these concepts and highlighting the promise of the platform, and address potential objections to dynamic benchmarking as a new standard for the field.

Rapid syntactic adaptation in self-paced reading: Detectable, but only with many participants.

Abstract: Temporarily ambiguous sentences that are disambiguated in favor of a less preferred parse are read more slowly than their unambiguous counterparts. This slowdown is referred to as a garden path effect. Recent self-paced reading studies have found that this effect decreased over the course of the experiment as participants were exposed to such syntactically ambiguous sentences. This decrease in the magnitude of the effect has been interpreted as evidence that readers calibrate their expectations to the context; this minimizes their surprise when they encounter these initially unexpected syntactic structures. Such recalibration of syntactic expectations, referred to as syntactic adaptation, is only one possible explanation for the decrease in garden path effect, however; this decrease could also be driven instead by increased familiarity with the self-paced reading paradigm (task adaptation). The goal of this article is to adjudicate between these two explanations. In a large between-group study (n = 642), we find evidence for syntactic adaptation over and above task adaptation. The magnitude of syntactic adaptation compared to task adaptation is very small, however. Power analyses show that a large number of participants is required to detect, with adequate power, syntactic adaptation in future between-subjects self-paced reading studies. This issue is exacerbated in experiments designed to detect modulations of the basic syntactic adaptation effect; such experiments are likely to be underpowered even with more than 1,200 participants. We conclude that while, contrary to recent suggestions, syntactic adaptation can be detected using self-paced reading, this paradigm is not very effective for studying this phenomenon. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

A Metaverse: Taxonomy, Components, Applications, and Open Challenges

Abstract: Unlike previous studies on the Metaverse based on Second Life, the current Metaverse is based on the social value of Generation Z that online and offline selves are not different. With the technological development of deep learning-based high-precision recognition models and natural generation models, Metaverse is being strengthened with various factors, from mobile-based always-on access to connectivity with reality using virtual currency. The integration of enhanced social activities and neural-net methods requires a new definition of Metaverse suitable for the present, different from the previous Metaverse. This paper divides the concepts and essential techniques necessary for realizing the Metaverse into three components (i.e., hardware, software, and contents) and three approaches (i.e., user interaction, implementation, and application) rather than marketing or hardware approach to conduct a comprehensive analysis. Furthermore, we describe essential methods based on three components and techniques to Metaverse’s representative Ready Player One, Roblox, and Facebook research in the domain of films, games, and studies. Finally, we summarize the limitations and directions for implementing the immersive Metaverse as social influences, constraints, and open challenges.

A Metaverse: Taxonomy, Components, Applications, and Open Challenges

Abstract: Unlike previous studies on the Metaverse based on Second Life, the current Metaverse is based on the social value of Generation Z that online and offline selves are not different. With the technological development of deep learning-based high-precision recognition models and natural generation models, Metaverse is being strengthened with various factors, from mobile-based always-on access to connectivity with reality using virtual currency. The integration of enhanced social activities and neural-net methods requires a new definition of Metaverse suitable for the present, different from the previous Metaverse. This paper divides the concepts and essential techniques necessary for realizing the Metaverse into three components (i.e., hardware, software, and contents) and three approaches (i.e., user interaction, implementation, and application) rather than marketing or hardware approach to conduct a comprehensive analysis. Furthermore, we describe essential methods based on three components and techniques to Metaverse’s representative Ready Player One, Roblox, and Facebook research in the domain of films, games, and studies. Finally, we summarize the limitations and directions for implementing the immersive Metaverse as social influences, constraints, and open challenges.

Holistic Evaluation of Language Models

Abstract: Language models (LMs) like GPT-3, PaLM, and ChatGPT are the foundation for almost all major language technologies, but their capabilities, limitations, and risks are not well understood. We present Holistic Evaluation of Language Models (HELM) to improve the transparency of LMs. LMs can serve many purposes and their behavior should satisfy many desiderata. To navigate the vast space of potential scenarios and metrics, we taxonomize the space and select representative subsets. We evaluate models on 16 core scenarios and 7 metrics, exposing important trade-offs. We supplement our core evaluation with seven targeted evaluations to deeply analyze specific aspects (including world knowledge, reasoning, regurgitation of copyrighted content, and generation of disinformation). We benchmark 30 LMs, from OpenAI, Microsoft, Google, Meta, Cohere, AI21 Labs, and others. Prior to HELM, models were evaluated on just 17.9% of the core HELM scenarios, with some prominent models not sharing a single scenario in common. We improve this to 96.0%: all 30 models are now benchmarked under the same standardized conditions. Our evaluation surfaces 25 top-level findings. For full transparency, we release all raw model prompts and completions publicly. HELM is a living benchmark for the community, continuously updated with new scenarios, metrics, and models https://crfm.stanford.edu/helm/latest/.

Hypothesis Only Baselines in Natural Language Inference

Abstract: We propose a hypothesis only baseline for diagnosing Natural Language Inference (NLI). Especially when an NLI dataset assumes inference is occurring based purely on the relationship between a context and a hypothesis, it follows that assessing entailment relations while ignoring the provided context is a degenerate solution. Yet, through experiments on ten distinct NLI datasets, we find that this approach, which we refer to as a hypothesis-only model, is able to significantly outperform a majority class baseline across a number of NLI datasets. Our analysis suggests that statistical irregularities may allow a model to perform NLI in some datasets beyond what should be achievable without access to the context.

PromptSource: An Integrated Development Environment and Repository for Natural Language Prompts

Abstract: PromptSource is a system for creating, sharing, and using natural language prompts. Prompts are functions that map an example from a dataset to a natural language input and target output. Using prompts to train and query language models is an emerging area in NLP that requires new tools that let users develop and refine these prompts collaboratively. PromptSource addresses the emergent challenges in this new setting with (1) a templating language for defining data-linked prompts, (2) an interface that lets users quickly iterate on prompt development by observing outputs of their prompts on many examples, and (3) a community-driven set of guidelines for contributing new prompts to a common pool. Over 2,000 prompts for roughly 170 datasets are already available in PromptSource. PromptSource is available at https://github.com/bigscience-workshop/promptsource.