Leading AI Experts Share Predictions for Machine Learning in 2020
Top machine learning minds like Google's Jeff Dean and PyTorch creator Soumith Chintala forecast major developments in natural language models and a growing emphasis on ethics beyond mere accuracy.
VentureBeat consults some of the most influential minds in artificial intelligence, including PyTorch creator Soumith Chintala, Google AI chief Jeff Dean, and IBM Research director Dario Gil, to share their predictions for machine learning in 2020. These experts highlight that AI is actively transforming the world today rather than at some vague point in the future.
The AI luminaries agree that 2019 sees significant strides in Transformer-based natural language models, and they expect this momentum to carry into the new year. Additionally, they anticipate ongoing controversy surrounding technologies like facial recognition and express a strong desire for the AI community to value more than just raw accuracy.
Chintala notes that while frameworks like PyTorch continue to grow rapidly in popularity and capabilities, the field lacks truly groundbreaking algorithmic breakthroughs. He points to the 2017 introduction of the Transformer architecture as the last major paradigm shift, suggesting that current progress relies heavily on refining and scaling existing methods.