Use of Artificial Intelligence Techniques to Improve the User Experience With Waste Recycling

In the wake of escalating global environmental challenges, sustainable waste management practices have emerged as a major concern. The need for efficient recycling strategies has been evolving. Recycling has become more and more popular and adopted amongst most of people lives. The need to find easier and more attractive way to encourage people to recycle more and more accurately. In this context, recycling practices with state-of-the-art natural language processing technologies presents an opportunity to improve the way we approach waste management. For the past 10 years, the technological advancements in Deep Natural Language Processing (DNLP) have been a huge milestone in the field of artificial intelligence. Since 2017, DNLP has progressed at rapid rate giving a leverage for a broader and more advanced applications. Everyday, new technologies, new models and papers give the opportunity to explore new original ways to solve various problems of all kinds. This paper embeds the development of a new recycling approach in a more intuitive way. Combining different state-of-the-art and recent models creating a speech-to-classification pipeline consisting of 2 different steps, a speech-to-text phase and then a classification phase. We create a new "user experience" of throwing wastes and recycling. This paper also presents the different steps of creating datasets to train and evaluate the models and the pipeline using different models, from text datasets in english and italian to audio datasets in italian.