Personally, I think that interpretability may be essential for alignability and debuggability, but that pure neural networks—ie those without some symbolic components—are unlikely to ever give us that. We may have to give up some performance to achieve that.
MACHINE LEARNING
-
Migrate Oracle to Lakehouse: Unlock Data Analytics and AI
By
–
Why should you migrate from Oracle to #Lakehouse? Because you can leverage the power of #data, analytics, and #AI in a single platform at scale and deliver business value faster. See how to convert PL/SQL proprietary code to an open stand Python code https://
dbricks.co/3Io5NvA -
Oxford’s RealFusion Enables 360° Object Reconstruction from Single Images
By
–
Oxford U Presents RealFusion: 360° Reconstructions of Any Object from a Single Image
-
Complex AI setup with CUDA, VS Studio, PyTorch, dependencies
By
–
AI is cool and all until you have to install CUDA, VS Studio for C++ support, PyTorch, and manage dependency versions with the 154 other libs.
-
ChatGPT as Python Assistant: Predicting Football Match Winners
By
–
KDnuggets News, January 25: ChatGPT as a Python Programming Assistant • Python and Machine Learning to Predict Football Match Winners: ChatGPT as a Python Programming Assistant • How to Use Python and Machine Learning to Predict Football Match Winners… https://
kdnuggets.com/2023/n03.html?
utm_source=dlvr.it&utm_medium=twitter&utm_campaign=kdnuggets-news-january-25-chatgpt-as-a-python-programming-assistant-python-and-machine-learning-to-predict-football-match-winners
… -
Scaled Q-Learning: Offline Reinforcement Learning Pre-Training Method
By
–
Presenting Scaled Q-Learning, a pre-training method for scaled offline #ReinforcementLearning that builds on the conservative Q-learning algorithm and enables efficient offline RL for multi-task training. Learn more → https://t.co/1lsHWcZFVM pic.twitter.com/aDq25su7V2
— Google AI (@GoogleAI) 23 février 2023Presenting Scaled Q-Learning, a pre-training method for scaled offline #ReinforcementLearning that builds on the conservative Q-learning algorithm and enables efficient offline RL for multi-task training. Learn more → https://
goo.gle/3XYA6hh -
AI and ML Healthcare Applications: Medical Imaging and Generative Models
By
–
Part 8 is out, posted by @greg_corrado & @ymatias on behalf of many, covering our work on AI and ML for healthcare-related applications, including medical imaging, ML in mobile healthcare settings, & the use of generative models like LLMs.
-
Data-Centric AI: Shifting Focus from Models to Datasets
By
–
AI systems involves data & model. Conventional machine learning focus on building better models. Data-Centric AI focus on curating datasets. Data-centric approach is an emerging field. So nice to see schools and orgs releasing course on this topic! https://
dcai.csail.mit.edu -
MIT 2023 Introduction to Data-Centric AI Course
By
–
[NEW COURSE] Introduction to Data-Centric AI – MIT 2023 A new course that covers many techniques related to data-centric approach such as data labelling, data curation, class imbalance, interpretability, data augmentation, prompt engineering… Lectures: https://
youtube.com/playlist?list=
PLnSYPjg2dHQKdig0vVbN-ZnEU0yNJ1mo5
…