Redirigiendo al acceso original de articulo en 24 segundos...
Inicio  /  Algorithms  /  Vol: 17 Par: 4 (2024)  /  Artículo
ARTÍCULO
TITULO

Uncertainty in Visual Generative AI

Kara Combs    
Adam Moyer and Trevor J. Bihl    

Resumen

Recently, generative artificial intelligence (GAI) has impressed the world with its ability to create text, images, and videos. However, there are still areas in which GAI produces undesirable or unintended results due to being ?uncertain?. Before wider use of AI-generated content, it is important to identify concepts where GAI is uncertain to ensure the usage thereof is ethical and to direct efforts for improvement. This study proposes a general pipeline to automatically quantify uncertainty within GAI. To measure uncertainty, the textual prompt to a text-to-image model is compared to captions supplied by four image-to-text models (GIT, BLIP, BLIP-2, and InstructBLIP). Its evaluation is based on machine translation metrics (BLEU, ROUGE, METEOR, and SPICE) and word embedding?s cosine similarity (Word2Vec, GloVe, FastText, DistilRoBERTa, MiniLM-6, and MiniLM-12). The generative AI models performed consistently across the metrics; however, the vector space models yielded the highest average similarity, close to 80%, which suggests more ideal and ?certain? results. Suggested future work includes identifying metrics that best align with a human baseline to ensure quality and consideration for more GAI models. The work within can be used to automatically identify concepts in which GAI is ?uncertain? to drive research aimed at increasing confidence in these areas.

 Artículos similares

       
 
Clara Pereira, Ana Silva, Cláudia Ferreira, Jorge de Brito, Inês Flores-Colen and José D. Silvestre    
In the field of building inspection and diagnosis, uncertainty is common and surveyors are aware of it, although it is not easily measured. This research proposes a model to quantify uncertainty based on the inspection of rendered façades. A Bayesian net... ver más
Revista: Infrastructures

 
Zhilin Lyu, Weitao Ding, Xiujun Sun, Hongqiang Sang, Ying Zhou, Peiyuan Yu and Lijun Zheng    
Aiming at the problems of difficult attitude stabilization, low landing accuracy, large external disturbance and slow dynamic response during the quadrotor dynamic landing on the wave glider, an improved series active disturbance rejection control method... ver más

 
Nikita Andriyanov    
The article is devoted to the study of convolutional neural network inference in the task of image processing under the influence of visual attacks. Attacks of four different types were considered: simple, involving the addition of white Gaussian noise, ... ver más
Revista: Applied Sciences

 
Bee Lan Oo, Teck Heng Benson Lim, Yixi Zhang    
Changes and challenges in employment are inevitable under the measures enacted to contain the COVID-19 pandemic. Early evidence suggests that the pandemic would disproportionately affect women compared to men. Focussing on women workforce in construction... ver más

 
Abdelrahman M. Abdallah, Rebecca A. Atadero and Mehmet E. Ozbek    
Bridge inspection standards in the United States require routine visual inspections to be conducted on most bridges at a maximum interval of two years regardless of the bridge condition. Limitations of this uniform calendar-based approach have been repor... ver más
Revista: Infrastructures