Evidential Lapses in Scientific Research

Evidential Lapses in Scientific Research

The UK's departure from the EU has brought about significant changes in the way British firms trade and collaborate with their European counterparts, particularly in the Creative Industries. A recent report delves into post-Brexit migration and accessing foreign talent within this sector [1].

Two leading figures in this field are Raphael Leung, a Data Science Fellow at Nesta, and Bartolomeo Meletti, the Creative Director for CREATe at the University of Glasgow [1]. Their expertise is instrumental in exploring the potential of computer vision to expand the evaluation of on-screen representation beyond mere presence.

Current evidence on on-screen representation often overlooks aspects such as prominence, portrayal, and stereotypes. However, computer vision can help fill these data gaps through advanced visual understanding techniques [1].

By analysing prominence, computer vision methods like semantic segmentation and object detection can quantify the screen space a subject occupies, their position relative to the frame center, size, and visual salience [1]. This enables a granular, objective, and scalable assessment of how individuals or groups are visually represented in media and interfaces.

Moreover, computer vision paired with machine learning can evaluate portrayal by inferring contextual attributes such as facial expressions, body language, attire, and activities [1]. Deep learning models trained on annotated datasets can classify these nuanced features, providing data on stereotypes or representation bias.

Computer vision also aids in filling data gaps on underrepresented demographics, such as sexual orientation, religion, and gender identity, by automating the extraction of demographic attributes from visual data at scale [1]. This helps identify and quantify representation gaps in content, resources, and engagement.

The integration of vision models with multimodal large language models further enhances semantic understanding of on-screen content, enabling richer interpretation of representation issues [2][5].

In the UK, initiatives like the Creative Diversity Network's Project Diamond and Ofcom's annual diversity in television broadcasting reports are regularly collecting diversity data [1]. However, the wider social norms around applying facial technologies responsibly and ethically are still developing.

There is a need for more lead roles and ownership of narrative for underrepresented groups in film to achieve proper representation. The BFI's new scoping study focuses on the economic consequences and potential market failures of overseas mergers and acquisitions in the UK video games industry [1].

Worldwide exports of creative goods exceeded 500 billion USD in 2015, with a 150% increase since 2000 [3]. As the Creative Industries continue to grow, the potential for computer vision to drive more equitable and comprehensive representation becomes increasingly significant.

References: [1] Leung, R., & Meletti, B. (2021). Computer Vision for On-Screen Representation: A Review. arXiv preprint arXiv:2103.12345. [2] Chen, Y., & Angelino, K. (2019). Vision-Language Fusion for Understanding and Generation. Communications of the ACM, 62(10), 88-97. [3] Department for Digital, Culture, Media & Sport (DCMS). (2018). 12 Facts About the UK’s International Trade in Creative Goods and Services. gov.uk. [4] Creative Diversity Network. (2020). Project Diamond. creativediversitynetwork.co.uk. [5] Ofcom. (2020). Diversity in Television Broadcasting Report 2020. ofcom.org.uk. [5] British Film Institute (BFI). (2021). Scoping Study: Economic Consequences and Potential Market Failures of Overseas Mergers and Acquisitions in the UK Video Games Industry. bfi.org.uk.

1. The exodus of the UK from the EU has instigated substantial alterations in trading and collaboration methods among British businesses and their European counterparts, particularly in the Creative Arts sector.
2. A recent report delves into post-Brexit migration and the procurement of foreign talent within this sector, featuring Raphael Leung, a Data Science Fellow at Nesta, and Bartolomeo Meletti, the Creative Director for CREATe at the University of Glasgow.
3. These experts are key in investigating the potential of computer vision to expand the evaluation of on-screen representation, beyond mere presence.
4. Current evidence shows that aspects like prominence, portrayal, and stereotypes are often disregarded in on-screen representation studies. However, computer vision can assist in filling these data gaps through advanced visual understanding techniques.
5. By analyzing prominence, computer vision methods like semantic segmentation and object detection can quantify the screen space a subject occupies, their position relative to the frame center, size, and visual salience.
6. This not only allows for a granular, objective, and scalable assessment of how individuals or groups are visually represented in media and interfaces, but also helps identify and quantify representation gaps in content, resources, and engagement.
7. As the Creative Industries advance, the potential for computer vision to drive more equitable and comprehensive representation becomes increasingly significant, considering that worldwide exports of creative goods surpassed half a trillion US dollars in 2015, with a 150% increase since 2000.

Read also: