-
DrawL: Understanding the Effects of Non-Mainstream Dialects in Prompted Image Generation
Authors:
Joshua N. Williams,
Molly FitzMorris,
Osman Aka,
Sarah Laszlo
Abstract:
Text-to-image models are now easy to use and ubiquitous. However, prior work has found that they are prone to recapitulating harmful Western stereotypes. For example, requesting that a model generate an "African person and their house," may produce a person standing next to a straw hut. In this example, the word "African" is an explicit descriptor of the person that the prompt is seeking to depict…
▽ More
Text-to-image models are now easy to use and ubiquitous. However, prior work has found that they are prone to recapitulating harmful Western stereotypes. For example, requesting that a model generate an "African person and their house," may produce a person standing next to a straw hut. In this example, the word "African" is an explicit descriptor of the person that the prompt is seeking to depict. Here, we examine whether implicit markers, such as dialect, can also affect the portrayal of people in text-to-image outputs. We pair prompts in Mainstream American English with counterfactuals that express grammatical constructions found in dialects correlated with historically marginalized groups. We find that through minimal, syntax-only changes to prompts, we can systematically shift the skin tone and gender of people in the generated images. We conclude with a discussion of whether dialectic distribution shifts like this are harmful or are expected, possibly even desirable, model behavior.
△ Less
Submitted 8 May, 2024;
originally announced May 2024.
-
A Distributional Semantics Approach to Implicit Language Learning
Authors:
Dimitrios Alikaniotis,
John N. Williams
Abstract:
In the present paper we show that distributional information is particularly important when considering concept availability under implicit language learning conditions. Based on results from different behavioural experiments we argue that the implicit learnability of semantic regularities depends on the degree to which the relevant concept is reflected in language use. In our simulations, we trai…
▽ More
In the present paper we show that distributional information is particularly important when considering concept availability under implicit language learning conditions. Based on results from different behavioural experiments we argue that the implicit learnability of semantic regularities depends on the degree to which the relevant concept is reflected in language use. In our simulations, we train a Vector-Space model on either an English or a Chinese corpus and then feed the resulting representations to a feed-forward neural network. The task of the neural network was to find a map** between the word representations and the novel words. Using datasets from four behavioural experiments, which used different semantic manipulations, we were able to obtain learning patterns very similar to those obtained by humans.
△ Less
Submitted 29 June, 2016;
originally announced June 2016.
-
Fractional abundance and the ecology of community structure
Authors:
Colleen K. Kelly,
Michael G. Bowler,
Jeffrey B. Joy,
John N. Williams
Abstract:
The ecological principle of limiting similarity dictates that species similar in resource requirements will compete, with the superior eventually excluding the inferior competitor from the community. The observation that nonetheless apparently similar species comprise a significant proportion of the diversity in any given community has led to suggestions that competition may not in fact be an impo…
▽ More
The ecological principle of limiting similarity dictates that species similar in resource requirements will compete, with the superior eventually excluding the inferior competitor from the community. The observation that nonetheless apparently similar species comprise a significant proportion of the diversity in any given community has led to suggestions that competition may not in fact be an important regulator of community structure and assembly. Here we apply a recently introduced metric of species interaction, fractional (relative) abundance, to tree species of the tropical wet forest of Barro Colorado Island, Panamá, the particular community that inspired the original model of non-niche or 'neutral' community dynamics9. We show a distribution of fractional abundances between pairs of most closely related congeneric tree species differing from that expected of competitive exclusion, but also inconsistent with expectations of simple similarity, whether such species interchangeability (a fundamental requirement of neutrality) is inferred at the community or the pair level. Similar evidence from a strikingly different dry forest has been linked to the focused, stable competition of a temporal niche dynamic. Taken together with these earlier findings, the results reported here establish a potentially widespread and important role for species interaction in the diversity and maintenance of natural communities that must be considered when inferring process from pattern.
△ Less
Submitted 15 August, 2010;
originally announced August 2010.
