turkmath.org

Machine Learning and Artificial Intelligence has been a very hot research and development topic over the last 10 years. Both Computer Vision and Natural Language Processing have seen significant advancements in recent years, Computer Vision has more mature and established products, while Natural Language Processing products aiming at widely adoption are still in the early stages of development. Well-known Computer Vision products like smart surveillance cameras with the ability to detect and track objects in real-time are widely used for some years now. Regarding Natural Language Processing, the first product aiming general adoption has just been released, the so-called ChatGPT. In both Computer Vision and Natural Language Processing, though some important advances are coming from academia, the state-of- the art models have been developed by big companies. Many Computer Vision research papers were published alongside models’ architecture and optimal weights. This allows other researchers to conduct Transfer Learning, a technique that uses a state-of-the-art pre-trained model as a starting point. In this talk, transfer learning will be introduced using a Computer Vision model for image classification and some research examples will be discussed.