The article also discusses the use of transformers in ML, stating that their novelty and effectiveness will likely keep them in the spotlight for some time. It suggests that the ML community is currently focused on improving and expanding upon the last successful innovation, similar to a gradient descent approach in algorithm space. The article ends by expressing interest in hearing about other lesser-known developments in the field.
Key takeaways:
- There is a growing interest in the field of Machine Learning beyond LLMs, with Cynthia Rudin's work on explainable AI being highlighted.
- NeRFS, a new approach to 3D graphics that uses machine learning to position and color glowing orbs based on multi-angle camera shots, is being explored.
- There is a debate on whether NeRFS will be more effective than well-optimized polygon-based systems like Nanite+photogrammetry.
- Despite the interest in new approaches, transformers in machine learning are still considered new and effective, and are expected to dominate the field for some time.