-
What happens when generative AI models train recursively on each others' generated outputs?
Authors:
Hung Anh Vu,
Galen Reeves,
Emily Wenger
Abstract:
The internet is full of AI-generated content while also serving as a common source of training data for generative AI (genAI) models. This duality raises the possibility that future genAI models may be trained on other models' generated outputs. Prior work has studied consequences of models training on their own generated outputs, but limited work has considered what happens if models ingest conte…
▽ More
The internet is full of AI-generated content while also serving as a common source of training data for generative AI (genAI) models. This duality raises the possibility that future genAI models may be trained on other models' generated outputs. Prior work has studied consequences of models training on their own generated outputs, but limited work has considered what happens if models ingest content produced by other models. Given society's increasing dependence on genAI tools, understanding downstream effects of such data-mediated model interactions is critical. To this end, we provide empirical evidence for how data-mediated interactions might unfold in practice, develop a theoretical model for this interactive training process, and show experimentally possible long-term results of such interactions. We find that data-mediated interactions can benefit models by exposing them to novel concepts perhaps missed in original training data, but also can homogenize their performance on shared tasks.
△ Less
Submitted 3 June, 2025; v1 submitted 27 May, 2025;
originally announced May 2025.
-
Exploring Causes of Representational Similarity in Machine Learning Models
Authors:
Zeyu Michael Li,
Hung Anh Vu,
Damilola Awofisayo,
Emily Wenger
Abstract:
Numerous works have noted significant similarities in how machine learning models represent the world, even across modalities. Although much effort has been devoted to uncovering properties and metrics on which these models align, surprisingly little work has explored causes of this similarity. To advance this line of inquiry, this work explores how two possible causal factors -- dataset overlap a…
▽ More
Numerous works have noted significant similarities in how machine learning models represent the world, even across modalities. Although much effort has been devoted to uncovering properties and metrics on which these models align, surprisingly little work has explored causes of this similarity. To advance this line of inquiry, this work explores how two possible causal factors -- dataset overlap and task overlap -- influence downstream model similarity. The exploration of dataset overlap is motivated by the reality that large-scale generative AI models are often trained on overlapping datasets of scraped internet data, while the exploration of task overlap seeks to substantiate claims from a recent work, the Platonic Representation Hypothesis, that task similarity may drive model similarity. We evaluate the effects of both factors through a broad set of experiments. We find that both positively correlate with higher representational similarity and that combining them provides the strongest effect. Our code and dataset are published.
△ Less
Submitted 20 May, 2025;
originally announced May 2025.
-
Integrating Preprocessing Methods and Convolutional Neural Networks for Effective Tumor Detection in Medical Imaging
Authors:
Ha Anh Vu
Abstract:
This research presents a machine-learning approach for tumor detection in medical images using convolutional neural networks (CNNs). The study focuses on preprocessing techniques to enhance image features relevant to tumor detection, followed by developing and training a CNN model for accurate classification. Various image processing techniques, including Gaussian smoothing, bilateral filtering, a…
▽ More
This research presents a machine-learning approach for tumor detection in medical images using convolutional neural networks (CNNs). The study focuses on preprocessing techniques to enhance image features relevant to tumor detection, followed by developing and training a CNN model for accurate classification. Various image processing techniques, including Gaussian smoothing, bilateral filtering, and K-means clustering, are employed to preprocess the input images and highlight tumor regions. The CNN model is trained and evaluated on a dataset of medical images, with augmentation and data generators utilized to enhance model generalization. Experimental results demonstrate the effectiveness of the proposed approach in accurately detecting tumors in medical images, paving the way for improved diagnostic tools in healthcare.
△ Less
Submitted 25 February, 2024;
originally announced February 2024.
-
Fast & Fair: Efficient Second-Order Robust Optimization for Fairness in Machine Learning
Authors:
Allen Minch,
Hung Anh Vu,
Anne Marie Warren
Abstract:
This project explores adversarial training techniques to develop fairer Deep Neural Networks (DNNs) to mitigate the inherent bias they are known to exhibit. DNNs are susceptible to inheriting bias with respect to sensitive attributes such as race and gender, which can lead to life-altering outcomes (e.g., demographic bias in facial recognition software used to arrest a suspect). We propose a robus…
▽ More
This project explores adversarial training techniques to develop fairer Deep Neural Networks (DNNs) to mitigate the inherent bias they are known to exhibit. DNNs are susceptible to inheriting bias with respect to sensitive attributes such as race and gender, which can lead to life-altering outcomes (e.g., demographic bias in facial recognition software used to arrest a suspect). We propose a robust optimization problem, which we demonstrate can improve fairness in several datasets, both synthetic and real-world, using an affine linear model. Leveraging second order information, we are able to find a solution to our optimization problem more efficiently than a purely first order method.
△ Less
Submitted 3 January, 2024;
originally announced January 2024.
-
Surjective Span 6 Cellular Automata
Authors:
Hung Anh Vu,
Nate Schnitzer,
Ethan Ewing
Abstract:
Using FSA and the construction algorithm, we generated a list of surjective span 6 cellular automata as a modest sample for our FDense program. We wanted to experimentally quantify Mike Boyle's conjecture which states that the jointly periodic points of one-dimensional cellular automata are dense. Furthermore, we wanted to know if the cardinality of cellular automata on N symbols is greater than o…
▽ More
Using FSA and the construction algorithm, we generated a list of surjective span 6 cellular automata as a modest sample for our FDense program. We wanted to experimentally quantify Mike Boyle's conjecture which states that the jointly periodic points of one-dimensional cellular automata are dense. Furthermore, we wanted to know if the cardinality of cellular automata on N symbols is greater than or equal to the square root of N.
△ Less
Submitted 10 June, 2023;
originally announced June 2023.
-
Intent expression using eye robot for mascot robot system
Authors:
Yoichi Yamazaki,
Fangyan Dong,
Yuta Masuda,
Yukiko Uehara,
Petar Kormushev,
Hai An Vu,
Phuc Quang Le,
Kaoru Hirota
Abstract:
An intent expression system using eye robots is proposed for a mascot robot system from a viewpoint of humatronics. The eye robot aims at providing a basic interface method for an information terminal robot system. To achieve better understanding of the displayed information, the importance and the degree of certainty of the information should be communicated along with the main content. The pro…
▽ More
An intent expression system using eye robots is proposed for a mascot robot system from a viewpoint of humatronics. The eye robot aims at providing a basic interface method for an information terminal robot system. To achieve better understanding of the displayed information, the importance and the degree of certainty of the information should be communicated along with the main content. The proposed intent expression system aims at conveying this additional information using the eye robot system. Eye motions are represented as the states in a pleasure-arousal space model. Changes in the model state are calculated by fuzzy inference according to the importance and degree of certainty of the displayed information. These changes influence the arousal-sleep coordinates in the space that corresponds to levels of liveliness during communication. The eye robot provides a basic interface for the mascot robot system that is easy to be understood as an information terminal for home environments in a humatronics society.
△ Less
Submitted 9 April, 2009;
originally announced April 2009.
-
Fuzzy inference based mentality estimation for eye robot agent
Authors:
Yoichi Yamazaki,
Fangyan Dong,
Yuta Masuda,
Yukiko Uehara,
Petar Kormushev,
Hai An Vu,
Phuc Quang Le,
Kaoru Hirota
Abstract:
Household robots need to communicate with human beings in a friendly fashion. To achieve better understanding of displayed information, an importance and a certainty of the information should be communicated together with the main information. The proposed intent expression system aims to convey this additional information using an eye robot. The eye motions are represented as states in a pleasu…
▽ More
Household robots need to communicate with human beings in a friendly fashion. To achieve better understanding of displayed information, an importance and a certainty of the information should be communicated together with the main information. The proposed intent expression system aims to convey this additional information using an eye robot. The eye motions are represented as states in a pleasure-arousal space model. Change of the model state is calculated by fuzzy inference according to the importance and certainty of the displayed information. This change influences the arousal-sleep coordinate in the space which corresponds to activeness in communication. The eye robot provides a basic interface for the mascot robot system which is an easy to understand information terminal for home environments in a humatronics society.
△ Less
Submitted 9 April, 2009;
originally announced April 2009.
