NotebookLM: The podcast we didn’t know we needed.  

8

October

2024

No ratings yet.

Recently, while I was scrolling through my Tik Tok app before I went to sleep, I came across a very intriguing video of a woman trying to explain the insights of a new science study. As a physics enthusiast I stopped and watched the whole video. Basically, the study was based on the origin and behavior of plasma, one of the four fundamental origins of matter. As of what I understood, what the researchers were proposing was that plasma could be a form of life and be responsible for some unexplained UFO phenomena, while it also seems to have caused some aircraft disaster incidents, based on astronaut disclosures. However, since the topic was already too hard to conceive, I decided at that moment to turn to NotebookLM and request help. 

NotebookLM is the new Google’s Gen AI tool that, similar to ChatGPT, can summarize any text we feed it with and generate instant insights. Nevertheless, it does not use the Internet’s public data as supplementary input, like most GenAI tools do, but instead only focuses on the data the user has loaded and comes with some predefined questions for better guidance. But this is not its most exciting feature. NotebookLM introduced a podcast feature in September, consisting of two speakers developing a conversation upon the topics addressed in the respective article or information given. The objective is to break down the topic in the simplest way for the listener to get the main idea.  

Back to my experience, after loading the article of the research to the platform, I requested for the podcast to be generated. It was a twelve-minute conversation between a man and a woman talking about what plasma is, how it’s connected to unexplaned phenomena, how it changes shape and the fact that it has curiosity: a basic human instict. The scariest and simultaneously most fascinating thing was their humanlike way of communication. Their voices had nothing to do with Siri or Alexa, but they were convincing enough to make me believe they were actual human beings.

I was pleasantly surprised when I realized I could indeed understand better all the components of the paper. They dived into every single aspect addressed and analyzed briefly all the important parts one would need to gain a high-level knowledge on the matter. The fact that the conversation was developed in a casual way accompanied by humoristic additions was what kept me interested, since I felt like I was also a part of it. I can surely admit that listening to this made the whole process of learning more entertaining.

This quick experience I had with the tool made me think of its potential. In my case, I can already see myself using it for getting a summarized version of the papers I will need to work on for my academic assignments. And besides the podcast feature, it can also optimize notes, divide the paper into its sub-topics, make new idea suggestions and even cite the sources when it adds information (Infosys, 2024). But I recommend it for more than just studying purposes. After all, sometimes we simply need to heal our curiosity about complicated matters, like plasma activity.  

References

Limited, I. (n.d.). NoteBook LM – powered by Gen AI. https://blogs.infosys.com/digital-experience/mobility/notebook-lm-powered-by-gen-ai.html

Please rate this

Intelligent Fashion; Can AI redefine Fashion Design?

17

September

2024

No ratings yet.

Fashion trends are an evolving inspiration and tend to influence deeply both customers and designers. What is more, as discussed in class, the level of personalization is a factor that can contribute highly to a person’s intention to buy. Now it is possible for designers to leverage generative AI systems to assist them in the designing process, while making sure that their work is always aligned with their customers’ needs. 

The fashion design process traditionally includes the designer gathering inspiration and illustrating creative ideas on paper. By the time a new clothing piece was launched, a long process of sketching, coloring, prototyping, adjusting, and fitting had taken place. However, AI-driven designing models are said to reduce costs by over 20%, introducing a breakthrough in the fashion industry [3]. 

Designers gain inspiration from the environment, art, culture and many more aspects of life and they tend to visualize their ideas by using images or even texts. Nonetheless, using the Internet for brainstorming can be quite challenging, since the search engines also return irrelevant results and also designers find it hard to transfer on paper the exact idea living in their minds. Using the appropriate model, all this unstructured raw data of media (image, text, video) can be guided and illustrated, even as sketches, and later be optimized by the personal touch of the respective creator [5]. Besides illustrating the initial idea, it is possible to combine dynamically different colors, patterns or textures in order to experiment with a variety of styles until they make final decisions. Generally, the automatic classification of fashion elements does not require designers to intervene much manually, enhancing productivity overall [1].  

This year HAIGEN (Human-AI Collaboration for GENeration) was introduced, a model developed to aid the work of designers, helping them tailor their patterns with the assistance of “sketch-to-image”, “image-to-sketch” and “text-to-image” models (some located locally and others cloud-based). These models can be personalized based on the way of working of each user and offer a private environment that helps boost the creativity of the creators and streamline the whole process [4].  

But the power of AI does not stop there. Luxury brands can now create pieces exclusively for their long-lasting customers or public features for marketing purposes. Designers can focus on sewing one-of-a-kind pieces and apply AI suggestions based on body measurements and proportions [2]. Real-time length adjustments and fabric or pattern combinations are also possible, as well as virtual try-ons to ensure that the final product will be fully personalized.  

Image1: Wearify: The Intersection of Fashion and Technology: How Generative AI is Changing the Game 

I find this use of AI to be particularly innovative. I have tried virtual try-ons in the past and I could really see the potential in them. More specifically, I tried Nike’s virtual representation of sport shoes and outfits, which is not yet close to the level of personalization discussed above, but still proposes a smart way to understand the product better. This is how I visualize designers to also follow the same path in an attempt to bring their ideas into reality and also customize them for their loyal customers.  

Not only is the fashion industry already shifting from conventional ways of working, but also customers will soon be able to experience for themselves the benefits technology can achieve. Data-driven solutions for fashion design optimization seem like they could solve fundamental problems designers face, especially when dealing with today’s high demand and lack of inspiration resources. After all, generative AI has proven to be a great ally in creativity! 

List of references:

[1] AI Assisted Fashion Design: a review. (2023). IEEE Journals & Magazine | IEEE Xplore. https://ieeexplore.ieee.org/abstract/document/10223039 

[2] Gong, W., & Khalid, L. (2021, January 20). Aesthetics, Personalization and Recommendation: A survey on Deep Learning in Fashion. arXiv.org. https://arxiv.org/abs/2101.08301 

[3] Harreis, H., Koullias, T., Roberts, R., & Te, K. (2023, March 8). Generative AI: Unlocking the future of fashion. McKinsey & Company. https://www.mckinsey.com/industries/retail/our-insights/generative-ai-unlocking-the-future-of-fashion 

[4] Jiang, J., Wu, D., Deng, H., Long, Y., Tang, W., Li, X., Liu, C., Jin, Z., Zhang, W., & Qi, T. (2024). HAIGEN: towards Human-AI Collaboration for facilitating creativity and style generation in fashion design. Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies, 8(3), 1–27. https://doi.org/10.1145/3678518 

[5] Tsipursky, G. (2023, November 21). Can AI outperform human creativity? Research suggests machines are more capable than we thought. Entrepreneur. https://www.entrepreneur.com/technology/can-ai-outperform-human-creativity-research-suggests/464485 

[6] Wearify. (2023, April 3). The Intersection of Fashion and Technology: How Generative AI is Changing the Game. https://www.linkedin.com/pulse/intersection-fashion-technology-how-generative-ai-changing/

Please rate this