Don't want to work from home? GAN's father, Ian Goodfellow, left Apple in anger

Reporting by XinZhiyuan

Editor: David Layan is sleepy

Due to the reluctance to implement Apple's "return to the office" policy, GAN's father Ian Goodfellow chose to leave his job.

Three years after joining the company, Apple's director of machine learning, Ian Goodfellow, abruptly announced his departure.

The reason behind this is somewhat unexpected: reluctance to return to the office.

Office, or remote is good

In a memo to employees, Ian Goodfellow wrote, "I firmly believe that more flexibility is the best policy for my team."

After the outbreak of the new crown epidemic, Apple embarked on a two-year remote work policy.

Since April 11, Apple employees have begun to return to the office one after another. Apple has taken a phased approach to its recovery work plan. At first, employees were required to be physically present at least one day a week. On May 4, Apple raised the requirement to two days a week.

Starting May 23, employees are required to work in the office three days a week. This is the beginning of Apple's so-called "hybrid" work program, which will require employees to work in the office every Monday, Tuesday and Thursday.

This plan to gradually increase the proportion of office work is currently met with considerable opposition. Not all Apple employees are keen on this initiative. A survey of a small number of employees found that a high percentage of employees are actively looking for other jobs.

Some employees believe that the policy may increase the likelihood of contracting COVID-19, deepen bad workplace culture, and disrupt work-life balance. These factors are all potential reasons for employee departures.

Under the social media reporting Goodfellow's departure, it was revealed that Apple's managers and executives are currently lobbying internally to try to get the company to retract its current "back to office" policy, but if the lobbying fails, these people may also face the fate of leaving.

On the other hand, not all big factories force employees to return to the office like Apple.

Goodfellow's former employer, Google, was much more flexible. Google has stipulated that some teams will resume working in person starting last month, but there are still many employees who can work from home for a long time.

Apple has also reportedly given individual teams some flexibility, allowing team leaders to adjust policies in the right way. But now it seems that this may not include Goodfellow's team.

While some Apple employees have previously chosen to leave due to the company's office policies, Goodfellow's departure is the most high-profile case reported publicly so far.

But so far, on Goodfellow's LinkedIn homepage, his job information has not been updated, and it still shows "Apple's head of machine learning."

The father of GAN, Bengio and Ng Enda's senior disciple

Who is Goodfellow? Judging from his resume, this man is really not ordinary.

Goodfellow is not only a computer scientist and engineer in his early Chinese New Year's Eve, but also a director at a young age.

He has served as a brain research scientist at Google and as director of machine learning at Apple.

The first thing that comes to mind when goodfellow comes to mind is his contributions to artificial neural networks and deep learning, and the famous generative adversarial network (GAN) was invented by him.

In fact, Goodfellow made GAN, really a bit of a fortuitous taste.

One night in 2014, he was celebrating with a recently graduated PhD student. After a while, a few of his friends asked him for help with a tricky project, and they were working on a computer program that would automatically generate pictures.

Address of the thesis: https://proceedings.neurips.cc/paper/2014/file/5ca3e9b122f61f8f06494c97b1afccf3-Paper.pdf

In fact, researchers have tried to use neural networks and algorithms, roughly imitating human brain neural networks, as generative models, from the main creator to create reasonable data.

However, the results of this operation are not very satisfactory. The resulting picture is often missing two pounds, either blurry or making the mistake of having fewer faces and ears.

Goodfellow's friend thought of a way to perform complex data analysis of the elements that make up the picture, so that the machine can better generate the picture.

Goodfellow pondered the question as he drank a beer. Suddenly he had an epiphany, why not let the two neural networks oppose?

His friends expressed skepticism. But he thinks it's a viable solution.

Hastily saying goodbye to friends, he returns home and discovers that his girlfriend is asleep, which gives him the opportunity to validate his ideas.

He immediately began experimenting with programming and then testing the software he had designed himself. Unexpectedly, he succeeded once.

This is the "GAN" in the beginning.

In addition to the name of the father of "GAN", he is also the author of the "Flower Book", which was originally called "Deep Learning". It can be said to be one of the must-read books for getting involved in the field of artificial intelligence.

Behind such a light resume is the same impeccable academic background.

Goodfellow finished his undergraduate and master's degrees in computer science at Stanford University, where he studied under AI giant Ng Enda.

During his Doctorate career, he studied at the University of Montreal in Canada under Yoshua Bengio and Aaron Courville, both of whom were well-known computer scientists.

Goodfellow's doctoral dissertation is closely related to deep learning and his computer vision applications, and it can be said that Goodfellow's interests and research focus have run through his entire academic career.

After graduating, Goodfellow first joined Google as part of its brain research team.

He then left Google to join the then-newly established OpenAI Research Lab. And in March 2017, he returned to Google.

Today, Goodfellow is already a well-known artificial intelligence expert. In 2017, he was selected by the MIT Science and Technology Review as one of the 35 innovators under the age of 35. In 2019, he was selected by the Foreign Policy as one of the 100 thinkers in the world.

That same year, he left Google again and joined Apple as the director of the machine learning special project group.

He has been working at Apple ever since.

People with ability have always been unique. This time Goodfellow protested the resignation of the company's policy, I don't know if it will force Apple's concessions.

After all, no one wants such a big cow to run to someone else.

Resources:

https://www.reddit.com/r/MachineLearning/comments/uks8zr/n_ian_goodfellow_apples_director_of_machine/

https://9to5mac.com/2022/05/08/apple-in-person-work-policy/

https://appleinsider.com/articles/22/05/07/apples-director-of-machine-learning-exits-over-return-to-office-policy