Are you curious what’s happened in the machine learning over the past week? We’ve got you covered! Today, we’re bringing you a (bi)weekly roundup with the latest news from the world of algorithms and artificial intelligence.
There’s a lot of interesting stories, so let’s get right to it!
Weekly Roundup: January 26 – February 8
As every week, on our blog, you can find professional articles on ML written by experts. We publish regularly so visit our blog for all the latest articles!
» Google Releases Eventarc into General Availability by Steef-Jan Wiggers on InfoQ | January 31
Eventarc is an eventing platform available on the Google Cloud Platform (GCP), allowing customers to send events to Cloud Run from more than 60 Google Cloud sources. In a recent blog post, the company announced the general availability (GA) of Eventarc.
In October last year, Google released the preview of Eventarc to provide customers with a service to connect Cloud Run services with events from various sources, adhering to the CloudEvents standard. With the GA release, the company made some updates.Check out what you can do with Eventarc.
» Unlimited computer fractals can help train AI to see by Will Douglas Heaven on MIT Technology Review | February 4
Researchers in Japan have shown that AIs can start learning to recognize everyday objects by being trained on computer-generated fractals instead. Read the article to learn why fractals are a good idea to train AI.
» ‘Weird new things are happening in software,’ says Stanford AI professor Chris Re by Tiernan Ray on ZDNet | January 31
Christopher Re, who is a Stanford University associate professor of computer science, gave a talk for the University’s Human-Centered Artificial Intelligence institute titled “Weird new things are happening in software.”
That weird new thing, is that the stuff that was important only a few years ago is now rather trivial, while new challenges are cropping up. But there is more to this story so make sure to read this piece!
» Elon Musk says his start-up Neuralink has wired up a monkey to play video games using its mind by Sam Shead on CNBC | February 1
Elon Musk said in an interview that a monkey has been wired up to play video games with its mind by a company he founded called Neuralink. Neuralink put a computer chip into the monkey’s skull and used “tiny wires” to connect it to its brain.
» Artificial intelligence must not be allowed to replace the imperfection of human empathy by Arshin Adib-Moghaddam on The Conversation | February 1
The author of this article states that “if AI research yields a new ideology centered around the notion of perfectionism and maximum productivity, then it will be a destructive force that will lead to more wars, more famines, and more social and economic distress, especially for the poor. At this juncture of global history, this choice is still ours.”
What do you think about it? This piece covers a ethics-oriented approach to AI.
» “Liquid” machine-learning system adapts to changing conditions by Daniel Ackerman on MIT News Office | January 28
The new type of neural network could aid decision making in autonomous driving and medical diagnosis.
MIT researchers have developed a type of neural network that learns on the job, not just during its training phase. These flexible algorithms, dubbed “liquid” networks, change their underlying equations to continuously adapt to new data inputs. The advance could aid decision making based on data streams that change over time, including those involved in medical diagnosis and autonomous driving.
» Here’s a Way to Learn if Facial Recognition Systems Used Your Photos by Cade Metz and Kashmir Hill on The New York Times | February 2
Let’s make it short – facial recognition is not as cool as you think when you’re on the other side of it. Next time you’re uploading your photos to the social media or other platfroms, think twice!
» This is how we lost control of our faces by Karen Hao on MIT Technology Review | February 5
The largest ever study of facial-recognition data shows how much the rise of deep learning has fueled a loss of privacy. And it’s serious.
That’s all folks! Have you found something of interest in this weekly roundup? We hope you got inspired! Don’t forget to check our blog for more inspiring articles.
And if you came across an interesting ML article, or maybe wrote one yourself and would like to share it with other people, let us know, we’ll spread the news in our weekly roundup! See you next time!
ML Experiment Tracking: What It Is, Why It Matters, and How to Implement It
10 mins read | Author Jakub Czakon | Updated July 14th, 2021
Let me share a story that I’ve heard too many times.
”… We were developing an ML model with my team, we ran a lot of experiments and got promising results…
…unfortunately, we couldn’t tell exactly what performed best because we forgot to save some model parameters and dataset versions…
…after a few weeks, we weren’t even sure what we have actually tried and we needed to re-run pretty much everything”
– unfortunate ML researcher.
And the truth is, when you develop ML models you will run a lot of experiments.
Those experiments may:
- use different models and model hyperparameters
- use different training or evaluation data,
- run different code (including this small change that you wanted to test quickly)
- run the same code in a different environment (not knowing which PyTorch or Tensorflow version was installed)
And as a result, they can produce completely different evaluation metrics.
Keeping track of all that information can very quickly become really hard. Especially if you want to organize and compare those experiments and feel confident that you know which setup produced the best result.
This is where ML experiment tracking comes in.Continue reading ->