Showing 1–10 of 10 results for author: Veselovsky, V

The AI Review Lottery: Widespread AI-Assisted Peer Reviews Boost Paper Scores and Acceptance Rates

Authors: Giuseppe Russo Latona, Manoel Horta Ribeiro, Tim R. Davidson, Veniamin Veselovsky, Robert West

Abstract: Journals and conferences worry that peer reviews assisted by artificial intelligence (AI), in particular, large language models (LLMs), may negatively influence the validity and fairness of the peer-review system, a cornerstone of modern science. In this work, we address this concern with a quasi-experimental study of the prevalence and impact of AI-assisted peer reviews in the context of the 2024… ▽ More Journals and conferences worry that peer reviews assisted by artificial intelligence (AI), in particular, large language models (LLMs), may negatively influence the validity and fairness of the peer-review system, a cornerstone of modern science. In this work, we address this concern with a quasi-experimental study of the prevalence and impact of AI-assisted peer reviews in the context of the 2024 International Conference on Learning Representations (ICLR), a large and prestigious machine-learning conference. Our contributions are threefold. Firstly, we obtain a lower bound for the prevalence of AI-assisted reviews at ICLR 2024 using the GPTZero LLM detector, estimating that at least $15.8\%$ of reviews were written with AI assistance. Secondly, we estimate the impact of AI-assisted reviews on submission scores. Considering pairs of reviews with different scores assigned to the same paper, we find that in $53.4\%$ of pairs the AI-assisted review scores higher than the human review ($p = 0.002$; relative difference in probability of scoring higher: $+14.4\%$ in favor of AI-assisted reviews). Thirdly, we assess the impact of receiving an AI-assisted peer review on submission acceptance. In a matched study, submissions near the acceptance threshold that received an AI-assisted peer review were $4.9$ percentage points ($p = 0.024$) more likely to be accepted than submissions that did not. Overall, we show that AI-assisted reviews are consequential to the peer-review process and offer a discussion on future implications of current trends △ Less

Submitted 3 May, 2024; originally announced May 2024.

Comments: Manoel Horta Ribeiro, Tim R. Davidson, and Veniamin Veselovsky contributed equally to this work

Do Llamas Work in English? On the Latent Language of Multilingual Transformers

Authors: Chris Wendler, Veniamin Veselovsky, Giovanni Monea, Robert West

Abstract: We ask whether multilingual language models trained on unbalanced, English-dominated corpora use English as an internal pivot language -- a question of key importance for understanding how language models function and the origins of linguistic bias. Focusing on the Llama-2 family of transformer models, our study uses carefully constructed non-English prompts with a unique correct single-token cont… ▽ More We ask whether multilingual language models trained on unbalanced, English-dominated corpora use English as an internal pivot language -- a question of key importance for understanding how language models function and the origins of linguistic bias. Focusing on the Llama-2 family of transformer models, our study uses carefully constructed non-English prompts with a unique correct single-token continuation. From layer to layer, transformers gradually map an input embedding of the final prompt token to an output embedding from which next-token probabilities are computed. Tracking intermediate embeddings through their high-dimensional space reveals three distinct phases, whereby intermediate embeddings (1) start far away from output token embeddings; (2) already allow for decoding a semantically correct next token in the middle layers, but give higher probability to its version in English than in the input language; (3) finally move into an input-language-specific region of the embedding space. We cast these results into a conceptual model where the three phases operate in "input space", "concept space", and "output space", respectively. Crucially, our evidence suggests that the abstract "concept space" lies closer to English than to other languages, which may have important consequences regarding the biases held by multilingual language models. △ Less

Submitted 8 June, 2024; v1 submitted 16 February, 2024; originally announced February 2024.

Comments: 12 pages. 28 with appendix

Evaluating Language Model Agency through Negotiations

Authors: Tim R. Davidson, Veniamin Veselovsky, Martin Josifoski, Maxime Peyrard, Antoine Bosselut, Michal Kosinski, Robert West

Abstract: We introduce an approach to evaluate language model (LM) agency using negotiation games. This approach better reflects real-world use cases and addresses some of the shortcomings of alternative LM benchmarks. Negotiation games enable us to study multi-turn, and cross-model interactions, modulate complexity, and side-step accidental evaluation data leakage. We use our approach to test six widely us… ▽ More We introduce an approach to evaluate language model (LM) agency using negotiation games. This approach better reflects real-world use cases and addresses some of the shortcomings of alternative LM benchmarks. Negotiation games enable us to study multi-turn, and cross-model interactions, modulate complexity, and side-step accidental evaluation data leakage. We use our approach to test six widely used and publicly accessible LMs, evaluating performance and alignment in both self-play and cross-play settings. Noteworthy findings include: (i) only closed-source models tested here were able to complete these tasks; (ii) cooperative bargaining games proved to be most challenging to the models; and (iii) even the most powerful models sometimes "lose" to weaker opponents △ Less

Submitted 16 March, 2024; v1 submitted 9 January, 2024; originally announced January 2024.

Comments: Accepted to ICLR 2024, code and link to project data are made available at https://github.com/epfl-dlab/LAMEN

Capturing Dynamics in Online Public Discourse: A Case Study of Universal Basic Income Discussions on Reddit

Authors: Rachel Kim, Veniamin Veselovsky, Ashton Anderson

Abstract: Societal change is often driven by shifts in public opinion. As citizens evolve in their norms, beliefs, and values, public policies change too. While traditional opinion polling and surveys can outline the broad strokes of whether public opinion on a particular topic is changing, they usually cannot capture the full multi-dimensional richness and diversity of opinion present in a large heterogene… ▽ More Societal change is often driven by shifts in public opinion. As citizens evolve in their norms, beliefs, and values, public policies change too. While traditional opinion polling and surveys can outline the broad strokes of whether public opinion on a particular topic is changing, they usually cannot capture the full multi-dimensional richness and diversity of opinion present in a large heterogeneous population. However, an increasing fraction of public discourse about public policy issues is now occurring on online platforms, which presents an opportunity to measure public opinion change at a qualitatively different scale of resolution and context. In this paper, we present a conceptual model of observed opinion change on online platforms and apply it to study public discourse on Universal Basic Income (UBI) on Reddit throughout its history. UBI is a periodic, no-strings-attached cash payment given to every citizen of a population. We study UBI as it is a clearly-defined policy proposal that has recently experienced a surge of interest through trends like automation and events like the COVID-19 pandemic. We find that overall stance towards UBI on Reddit significantly declined until mid-2019, when this historical trend suddenly reversed and Reddit became substantially more supportive. Using our model, we find the most significant drivers of this overall stance change were shifts within different user cohorts, within communities that represented similar affluence levels, and within communities that represented similar partisan leanings. Our method identifies nuanced social drivers of opinion change in the large-scale public discourse that now regularly occurs online, and could be applied to a broad set of other important issues and policies. △ Less

Submitted 15 December, 2023; originally announced December 2023.

Prevalence and prevention of large language model use in crowd work

Authors: Veniamin Veselovsky, Manoel Horta Ribeiro, Philip Cozzolino, Andrew Gordon, David Rothschild, Robert West

Abstract: We show that the use of large language models (LLMs) is prevalent among crowd workers, and that targeted mitigation strategies can significantly reduce, but not eliminate, LLM use. On a text summarization task where workers were not directed in any way regarding their LLM use, the estimated prevalence of LLM use was around 30%, but was reduced by about half by asking workers to not use LLMs and by… ▽ More We show that the use of large language models (LLMs) is prevalent among crowd workers, and that targeted mitigation strategies can significantly reduce, but not eliminate, LLM use. On a text summarization task where workers were not directed in any way regarding their LLM use, the estimated prevalence of LLM use was around 30%, but was reduced by about half by asking workers to not use LLMs and by raising the cost of using them, e.g., by disabling copy-pasting. Secondary analyses give further insight into LLM use and its prevention: LLM use yields high-quality but homogeneous responses, which may harm research concerned with human (rather than model) behavior and degrade future models trained with crowdsourced data. At the same time, preventing LLM use may be at odds with obtaining high-quality responses; e.g., when requesting workers not to use LLMs, summaries contained fewer keywords carrying essential information. Our estimates will likely change as LLMs increase in popularity or capabilities, and as norms around their usage change. Yet, understanding the co-evolution of LLM-based tools and users is key to maintaining the validity of research done using crowdsourcing, and we provide a critical baseline before widespread adoption ensues. △ Less

Submitted 24 October, 2023; originally announced October 2023.

Comments: VV and MHR equal contribution. 14 pages, 1 figure, 1 table

Tube2Vec: Social and Semantic Embeddings of YouTube Channels

Authors: Léopaul Boesinger, Manoel Horta Ribeiro, Veniamin Veselovsky, Robert West

Abstract: Research using YouTube data often explores social and semantic dimensions of channels and videos. Typically, analyses rely on laborious manual annotation of content and content creators, often found by low-recall methods such as keyword search. Here, we explore an alternative approach, using latent representations (embeddings) obtained via machine learning. Using a large dataset of YouTube links s… ▽ More Research using YouTube data often explores social and semantic dimensions of channels and videos. Typically, analyses rely on laborious manual annotation of content and content creators, often found by low-recall methods such as keyword search. Here, we explore an alternative approach, using latent representations (embeddings) obtained via machine learning. Using a large dataset of YouTube links shared on Reddit; we create embeddings that capture social sharing behavior, video metadata (title, description, etc.), and YouTube's video recommendations. We evaluate these embeddings using crowdsourcing and existing datasets, finding that recommendation embeddings excel at capturing both social and semantic dimensions, although social-sharing embeddings better correlate with existing partisan scores. We share embeddings capturing the social and semantic dimensions of 44,000 YouTube channels for the benefit of future research on YouTube: https://github.com/epfl-dlab/youtube-embeddings. △ Less

Submitted 29 June, 2023; originally announced June 2023.

Artificial Artificial Artificial Intelligence: Crowd Workers Widely Use Large Language Models for Text Production Tasks

Authors: Veniamin Veselovsky, Manoel Horta Ribeiro, Robert West

Abstract: Large language models (LLMs) are remarkable data annotators. They can be used to generate high-fidelity supervised training data, as well as survey and experimental data. With the widespread adoption of LLMs, human gold--standard annotations are key to understanding the capabilities of LLMs and the validity of their results. However, crowdsourcing, an important, inexpensive way to obtain human ann… ▽ More Large language models (LLMs) are remarkable data annotators. They can be used to generate high-fidelity supervised training data, as well as survey and experimental data. With the widespread adoption of LLMs, human gold--standard annotations are key to understanding the capabilities of LLMs and the validity of their results. However, crowdsourcing, an important, inexpensive way to obtain human annotations, may itself be impacted by LLMs, as crowd workers have financial incentives to use LLMs to increase their productivity and income. To investigate this concern, we conducted a case study on the prevalence of LLM usage by crowd workers. We reran an abstract summarization task from the literature on Amazon Mechanical Turk and, through a combination of keystroke detection and synthetic text classification, estimate that 33-46% of crowd workers used LLMs when completing the task. Although generalization to other, less LLM-friendly tasks is unclear, our results call for platforms, researchers, and crowd workers to find new ways to ensure that human data remain human, perhaps using the methodology proposed here as a step** stone. Code/data: https://github.com/epfl-dlab/GPTurk △ Less

Submitted 13 June, 2023; originally announced June 2023.

Comments: 9 pages, 4 figures

Generating Faithful Synthetic Data with Large Language Models: A Case Study in Computational Social Science

Authors: Veniamin Veselovsky, Manoel Horta Ribeiro, Akhil Arora, Martin Josifoski, Ashton Anderson, Robert West

Abstract: Large Language Models (LLMs) have democratized synthetic data generation, which in turn has the potential to simplify and broaden a wide gamut of NLP tasks. Here, we tackle a pervasive problem in synthetic data generation: its generative distribution often differs from the distribution of real-world data researchers care about (in other words, it is unfaithful). In a case study on sarcasm detectio… ▽ More Large Language Models (LLMs) have democratized synthetic data generation, which in turn has the potential to simplify and broaden a wide gamut of NLP tasks. Here, we tackle a pervasive problem in synthetic data generation: its generative distribution often differs from the distribution of real-world data researchers care about (in other words, it is unfaithful). In a case study on sarcasm detection, we study three strategies to increase the faithfulness of synthetic data: grounding, filtering, and taxonomy-based generation. We evaluate these strategies using the performance of classifiers trained with generated synthetic data on real-world data. While all three strategies improve the performance of classifiers, we find that grounding works best for the task at hand. As synthetic data generation plays an ever-increasing role in NLP research, we expect this work to be a step** stone in improving its utility. We conclude this paper with some recommendations on how to generate high(er)-fidelity synthetic data for specific tasks. △ Less

Submitted 24 May, 2023; originally announced May 2023.

Comments: 8 pages

Reddit in the Time of COVID

Authors: Veniamin Veselovsky, Ashton Anderson

Abstract: When the COVID-19 pandemic hit, much of life moved online. Platforms of all types reported surges of activity, and people remarked on the various important functions that online platforms suddenly fulfilled. However, researchers lack a rigorous understanding of the pandemic's impacts on social platforms, and whether they were temporary or long-lasting. We present a conceptual framework for studyin… ▽ More When the COVID-19 pandemic hit, much of life moved online. Platforms of all types reported surges of activity, and people remarked on the various important functions that online platforms suddenly fulfilled. However, researchers lack a rigorous understanding of the pandemic's impacts on social platforms, and whether they were temporary or long-lasting. We present a conceptual framework for studying the large-scale evolution of social platforms and apply it to the study of Reddit's history, with a particular focus on the COVID-19 pandemic. We study platform evolution through two key dimensions: structure vs. content and macro- vs. micro-level analysis. Structural signals help us quantify how much behavior changed, while content analysis clarifies exactly how it changed. Applying these at the macro-level illuminates platform-wide changes, while at the micro-level we study impacts on individual users. We illustrate the value of this approach by showing the extraordinary and ordinary changes Reddit went through during the pandemic. First, we show that typically when rapid growth occurs, it is driven by a few concentrated communities and within a narrow slice of language use. However, Reddit's growth throughout COVID-19 was spread across disparate communities and languages. Second, all groups were equally affected in their change of interest, but veteran users tended to invoke COVID-related language more than newer users. Third, the new wave of users that arrived following COVID-19 was fundamentally different from previous cohorts of new users in terms of interests, activity, and likelihood of staying active on the platform. These findings provide a more rigorous understanding of how an online platform changed during the global pandemic. △ Less

Submitted 21 April, 2023; originally announced April 2023.

Comments: 12 pages, published in ICWSM 2023

The Amplification Paradox in Recommender Systems

Authors: Manoel Horta Ribeiro, Veniamin Veselovsky, Robert West

Abstract: Automated audits of recommender systems found that blindly following recommendations leads users to increasingly partisan, conspiratorial, or false content. At the same time, studies using real user traces suggest that recommender systems are not the primary driver of attention toward extreme content; on the contrary, such content is mostly reached through other means, e.g., other websites. In thi… ▽ More Automated audits of recommender systems found that blindly following recommendations leads users to increasingly partisan, conspiratorial, or false content. At the same time, studies using real user traces suggest that recommender systems are not the primary driver of attention toward extreme content; on the contrary, such content is mostly reached through other means, e.g., other websites. In this paper, we explain the following apparent paradox: if the recommendation algorithm favors extreme content, why is it not driving its consumption? With a simple agent-based model where users attribute different utilities to items in the recommender system, we show through simulations that the collaborative-filtering nature of recommender systems and the nicheness of extreme content can resolve the apparent paradox: although blindly following recommendations would indeed lead users to niche content, users rarely consume niche content when given the option because it is of low utility to them, which can lead the recommender system to deamplify such content. Our results call for a nuanced interpretation of ``algorithmic amplification'' and highlight the importance of modeling the utility of content to users when auditing recommender systems. Code available: https://github.com/epfl-dlab/amplification_paradox. △ Less

Submitted 5 April, 2023; v1 submitted 22 February, 2023; originally announced February 2023.

Comments: Accepted at ICWSM'23 please cite accordingly