CodeIsGo.com

New research has exploded the space of problems that quantum computers can efficiently verify, simultaneously knocking down milestone problems in quantum physics and math. By Mordechai Rorvig.

If today’s computers had dreams and ambitions, there would be some problems that they wouldn’t even dream about solving. Some problems would take far too much time or memory, running on for nearly forever. But in an exciting new result, a team of computer scientists have shown how in theory quantum computers should be able to rapidly verify that a practically infinite problem was solved.

The authors show that quantum computers can rapidly verify the solution to what is known as the halting problem. What allows this remarkable feat is quantum entanglement: each quantum computer contains subatomic particles that are entangled with particles in the other computer.

Entanglement offers a profound communication resource for the two quantum computers. In separate work from 2019, two of the new study’s coauthors, Anand Natarajan and John Wright, showed that entanglement allowed rapid verification of the solution of one class of enormously long problems, called NEEXP problems. To learn more follow the link for truly fascinating reading!

Constantine Yurevich is author of this article in which he argues why you should create a custom attribution tool. With out-of-the-box tools, you’re limited by their functionality, data transformations, models, and heuristics.

With raw data, you can build any attribution model that fits your business and domain area best. While few small companies collect raw data, most large businesses bring it into a data warehouse and visualize it using BI tools such as Google Data Studio, Tableau, Microsoft Power BI, Looker, etc.

• The challenge of attribution
• What machine learning changes
• Four steps to build an attribution model with Google BigQuery ML
  • Step 1: Feature mining
  • Step 2: Training a model
  • Step 3: Model evaluation
  • Step 4: Building an attribution model
• Pros and cons of behavior-based attribution

Quite in depth information about attribution models in general and good steering for anybody interesting in creating a custom attribution tool. Well worth your time!

By adopting agile ways of working, startups as well as established companies have successfully brought customer-centric digital initiatives to market. By Lars Marquardt, Jan-Philipp Martini, Georg Kappen, and Carsten Schaetzberger.

Digital leaders are now being asked to repeat this success with more complex, business-centric, transformative initiatives.

The dilemma is this: Digital teams can use agile ways of working to build customer-centric solutions because they are vertically adjacent to the core processes. Transformative initiatives, however, run horizontally across the value chain and must integrate with the core. Digital teams, therefore, are less free to use agile ways of working with such initiatives because they cannot ignore existing corporate processes and systems.

To meet senior leaders’ high expectations and achieve horizontal digitalization, leaders need to combine the agile ways of working used by startups with traditional, more rigid, corporate ways of working. Excellent read!

Brinnae Bent and Jessilyn Dunn form Duke University wrote an article about wearable technology. Wearable technology has the potential to transform healthcare and research through accessible, continuous, and longitudinal health monitoring.

The accuracy of wearable technologies has been a hotly debated topic both in research and in popular culture. As wearable technologies are increasingly being used for clinical research and healthcare, it is critical to understand their accuracy and determine how measurement errors may affect research conclusions and impact healthcare decision-making.

Further, the type of physical activity going on during a measurement also plays an important role in the accuracy of the heart rate measurement: we saw an approximate 30% drop in accuracy when a person is walking versus when they are seated at rest

In a study, the authors tested optical heart rate sensors on six of the most popular devices on the market for both consumers (Apple Watch, Fitbit, Garmin, XiaoMi Miband) and for research (Empatica E4, Biovotion Everion) against the clinical reference standard for monitoring heart rate, the electrocardiogram (ECG).

Click the link to the full article to learn more. You will leanr how accurate data actuall wearable devices can provide. Nice one!

Jeff Hawkins has a principle that intuitively makes a lot of sense, yet is something that Deep Learning research has not emphasized enough. This is the notion of embodied learning. That is, biological systems learn from interacting with the environment. By Carlos E. Perez.

The classic Deep Learning training procedure is one of the crudest teaching methods that one can possibly imagine. It is based on repetitively and randomly presenting facts about the world and hoping that the student (i.e. the neural network) is able to disentangle and create sufficient abstractions of the world.

Whereas most previous work focuses on the static setting (e.g., with images), we postulate that some of the causal factors could be discovered if the learner is allowed to interact with its environment.

Read the article in full to get better idea about Artificial Intelligence, challenges with Deep learning. The links to further reading and white papers are also provided. Nice one!

We’ve tested all the top AMD and Intel processors to help you make the right choice. By Dave James.

The best CPU for gaming simply has to be the processor to offer the heady combination of powerful single-threaded performance, some tasty multi-threaded processing chops, and a price tag that doesn’t instantly nuke your bank balance. Thankfully that holy trinity is now more readily available that at any time we can remember.

Some of the processors (CPU) they’ve tested in the article:

• Amd Ryzen 5 3600
• Amd RyzeN 5 3600X
• Amd RyzeN 5 2600
• Intel Core i5 8400
• Intel Core I5 9600K
• Intel Core I9 9900K
• Amd Ryzen 9 3900X
• Amd Ryzen 7 3700X

… and others. Despite the dominance of multi-core CPUs in today’s gaming rigs, the difficulty in coding for chips with multiple processor cores has meant we’re still not seeing many modern game engines taking full advantage of the powerful CPUs many of us now have in our machines. In fact, since AMD has decided to take the fight to Intel in terms of core-count, even six cores is becoming more and more common. Good read!

An article by Ying Wang about his experience with trying Haskell after using Python in production for few years. Even as it’s a wonderful Swiss Army knife, Python also feels limiting in some ways. The same classes of bugs (e.g. type casting / translation errors, state management errors) kept cropping up, and Sisyphean bugs frustrate me. Python is also quite slow when you compare the operations you want to execute vs. the theoretical maximum performance of those operations on the underlying hardware. He kept wondering about what was possible if he had used a different tool.

Since Haskell is a lambda calculus, and everything is functional, I thought I could get parallelism trivially. Nope. As far as I can tell, you still have to mutate your source code, and use a library, such as Control.Parallel (external dep) or Control.Concurrent (base stdlib). Simon Marlow’s “Parallel and Concurrent Programming in Haskell” is probably a must-read to understand this topic (and things like Control.Concurrent.MVar).

His experience is divided into the sections:

• What I like about Haskell
  • The type system
  • An emphasis on structure
  • Streams and stream-oriented programming foundations
  • Property-based testing
  • Hoogle
• What I dislike about Haskell
  • Cognitive load
  • Indentation
  • Concurrency / Parallelism
  • Package Management
  • Strictness / Non-strictness
  • Compatibility

… and much more. One thing to point out is Haskell isn’t a magical elixir. It’s still a programming language. Good read!

A massive neural network connects cameras, a robot arm, and a suction gripper in Covariant’s logistics system. By Evan Ackerman. Covariant.ai spent almost a year talking with literally hundreds of different companies about how smarter robots could potentially make a difference for them.

These companies represent sectors that include electronics manufacturing, car manufacturing, textiles, bio labs, construction, farming, hotels, elder care—“pretty much anything you could think about where maybe a robot could be helpful,” Pieter Abbeel tells us.

“People often say neural networks are just black boxes and if you’re doing something new you have to start from scratch. That’s not really true . . . Their strength comes from the fact that you can train end-to-end, you can train from input to the desired output”.

This is a very interesting interview for anybody into automation, neural networks, robotics and deep learning. Excellent!

Jesse Howarth wrote this piece about how they seen success with Rust on the client side and server side. For example, they use it on the client side for our video encoding pipeline for Go Live and on the server side for Elixir NIFs. Most recently, they drastically improved the performance of a service by switching its implementation from Go to Rust.

Rust is blazingly fast and memory-efficient: with no runtime or garbage collector, it can power performance-critical services, run on embedded devices, and easily integrate with other languages.

Their post explains why it made sense for us to reimplement the service, how it was done, and the resulting performance improvements:

• The Read States service
• Why Go did not meet our performance targets
• So why 2 minute spikes?
• Memory management in Rust
• Async Rust
• Implementation, load testing, and launch
• Raising the cache capacity

To explain why Go wasn’t meeting their performance targets, they also discuss the data structures, scale, access patterns, and architecture of the service. Good read!

Responsive web design has been a must-have skill for front-end developers for many years now, but creating a truly fluid interface and giving users a seamless experience across any device still poses a lot of challenges for designers and developers alike. By Razvan Onofrei.

The web is a wonderful medium, buzzing with information and connections. The web is also a very dynamic, interactive medium that, unlike the printed page, can and will be accessed from all sorts of devices. Web users get to easily adapt the web to their daily lives, making it part of their existence.

The article main sections are:

• A brief history, considerations for various patterns for solving different layout or UX challenges
• The first steps forward: breakpoints and media queries
• Giving it a shot: intermediate breakpoints
• Responsive typography
• Pushing the boundaries
• We found a solution that fits our needs

Plenty of useful information for anybody invested in the front design and typography. Nice one!