Deep Learning Advances to Early Adopter Phase in 2021 AI Trends
The August 2021 InfoQ trends report highlights deep learning's transition into the early adopter category and notes emerging challenges in edge deployment. The report also covers growing automation in machine learning pipelines and the steady evolution of commercial robotics platforms.
The latest InfoQ trends report reveals that deep learning officially transitions from the innovator stage to the early adopter category in 2021. This shift occurs as major frameworks like TensorFlow and PyTorch see widespread industry adoption. However, this progress brings new technical challenges, specifically regarding the deployment of complex algorithms onto edge devices and the management of exceptionally large training models.
Machine learning operations experience significant improvements as Kubernetes makes it easier to deploy ML models within standard compute stacks. The industry also sees a surge in tools that automate crucial pipeline steps, including data collection and model retraining. Additionally, AutoML emerges as a promising technology that frees data scientists from tedious hyperparameter optimization so they can focus on solving actual domain problems.
Beyond core deep learning, the report identifies GPU programming as a highly promising yet underutilized technology with potential for broader applications. Commercial robotics platforms slowly gain traction outside of academic circles, though experts believe many industry use cases remain undiscovered. As these technologies mature, software engineers and data scientists face an evolving landscape that requires continuous adaptation and skill development.