CodeIsGo.com

Recent advancements in deep reinforcement learning (deep RL) has enabled legged robots to learn many agile skills through automated environment interactions. In the past few years, researchers have greatly improved sample efficiency by using off-policy data, imitating animal behaviors, or performing meta learning. Posted by Yuxiang Yang and Deepali Jain, AI Residents, Robotics at Google.

However, sample efficiency remains a bottleneck for most deep reinforcement learning algorithms, especially in the legged locomotion domain.

The authors present two projects that aim to address the above problems and help close the perception-actuation loop for legged robots.

• In Data Efficient Reinforcement Learning for Legged Robots, they present an efficient way to learn low level motion control policies.
• Going beyond simple behaviors, we explore automatic path navigation in Hierarchical Reinforcement Learning for Quadruped Locomotion

Reinforcement learning poses a promising future for robotics by automating the controller design process. With model-based RL, we enabled efficient learning of generalizable locomotion behaviors directly on the real robot. With hierarchical RL, the robot learned to coordinate policies at different levels to achieve more complex tasks. In the future, we plan to bring perception into the loop, so that robots can operate truly autonomously in the real world.

Follow the link to full article to learn more. Very exciting!

At the foundation of clustered systems are so-called membership protocols. The job of a membership protocol is to keep clustered applications up-to-date with the list of nodes that are part of the cluster, allowing all the individual nodes to act as one system. By Manuel Bernhardt.

One of the motivations for Rapid was to be faster at scale than existing consensus protocols. The Rapid paper shows that Rapid can form a 2000 node cluster 2-5.8 times faster than Memberlist (Hashicor’s Lifeguard implementation) and Zookeeper. One of the problems with membership services that rely entirely on random gossip is that random gossip leads to higher tail latencies for convergence.

Rapid is designed around the central idea that a scalable membership service needs to have high confidence in failures before acting on them and that membership change decisions affecting multiple nodes should be grouped as opposed to happen on a per-case basis. Dissemination of membership information in Rapid happens via broadcast.

The article then guides you through:

• The Rapid membership service
• Dissemination
• Consensus
• Failure detection
• Joining the cluster
• What Rapid brings to the table

Another one of the big contrasts that in author’s opinion sets Rapid ahead of other membership protocols is the strong stability provided by the multi-node cut detector. A flaky node can be quite problematic in protocols where the failure detection mechanism doesn’t provide this type of confidence. Super interesting read!

The Angular framework was designed with flexibility in mind. That approach allows Angular applications to be executed across different environments — browser, server, web-worker, and even mobile devices are possible. Written by Nikita Poltoratsky, a Solution Architect at Akveo, open-source contributor and tech author.

In this series of articles, author reveals to you how does it even possible – execute Angular applications across different environments. Also, we will learn how to build custom Angular platform which renders applications inside the system’s terminal using ASCII graphics.

The article is split into these parts:

• Renderer
• Sanitizer
• Error handling
• Terminal module
• Platform terminal
• Building terminal application

You will get a custom platform creation process overview and learn about crucial Angular services and modules. Finally, you will build custom platform which renders Angular applications inside the system’s terminal using ASCII graphics!

And also yoou will get access to the repository with all the source files related to the Terminal platform.. Nice one!

Darren Gipson wrote this piece about their effort to standardize the Puppet code design. They have moved as much of the data layer as possible to Hiera. Hiera is a built-in key-value configuration data lookup system, used for separating data from Puppet code.

Moving the data layer to Hiera means that we only have to understand a simple yaml file when building a new environment. This allows for great code reuse and helps explain to others not familiar with our Puppet code where the settings are.

Hiera was the clear solution for their problem because it provides you with an entry point to store data that you can then read with Puppet code.

The article discusses:

• Business problem – migration to cloud-based services,
• Solution
• Complex examples

Using hashes in Hiera in certain way allows your Puppet code to be very data-driven, preventing the need to write the same code twice and providing flexibility for different configurations across environments. For details and code example follow the link to full article. Very good!

Often when we talk about accessibility problems, we end up talking about a number of different errors that all boil down to a missing accessible name: form fields without labels, images without alts, icon buttons without readable text, and many more. Written by Sarah Higley.

A missing or incorrect accessible name in some form or other is right up there with poor color contrast in the list of most common accessibility errors across the web.

More recently, a greater awareness of accessibility and an increase in the use of ARIA attributes has resulted in a sort of reverse naming problem, where elements that cannot or should not be named get artificial names through ARIA. While an improperly added name is a step up from a missing one, it will often give the illusion of accessibility while masking an underlying problem that remains unaddressed.

Here’s a non-comprehensive list of terms and concepts that all boil down to talking about names:

• labels, the element, and associating labels with form fields
• table captions
• image alt text
• the element
• an SVG element
• aria-label and aria-labelledby

Not every element can support an accessible name, and even among those that can, not every one should. To learn about few general categories of elements that should always be named read the article in full. Excellent!

The old drug discovery system was built to benefit shareholders, not patients. But a new, Linux-like platform could transform the way medicine is developed—and energize the race against COVID-19. By Ruth Reader, writer for Fast Company.

The response to COVID-19 has been more open-source than any drug effort in modern memory. On January 11, less than two weeks after the virus was reported to the World Health Organization, Chinese researchers published a draft of the virus’s genetic sequence. The information enabled scientists across the globe to begin developing tests, treatments, and vaccines. Pharmaceutical companies searched their archives for drugs that might be repurposed as treatments for COVID-19 and formed consortiums to combine resources and expedite the process. These efforts have yielded some 90 vaccine candidates, seven of which are in Phase I trials and three of which are advancing to Phase II. There are nearly 1,000 clinical trials listed with the Centers for Disease Control and Prevention related to COVID-19.

In recent history, pharmaceutical companies have failed to deliver treatments for devastating illnesses because they cannot easily profit from them. While they may investigate diseases for which there is no treatment or cure and even find compelling insights, they won’t necessarily turn those findings into medicine.

A human rights lawyer named Jaykumar Menon is now building out a platform where scientists and researchers can freely access technological tools for researching disease, share their discoveries, launch investigations into molecules or potential drugs, and find entities to turn that research into medicine. If the platform succeeds, it would allow drugs to succeed on their merit and need, rather than their ability to be profitable.

Super interesting reading. Highly recommended!

Accessing a locked system is always a challenge. Full-disk encryption presents an immediate challenge to forensic experts. When acquiring computers with encrypted system volumes, the investigation cannot go forward without breaking the encryption first. By Oleg Afonin from ElcomSoft.

Traditionally, experts would remove the hard drive(s), make disk images and work from there. There is a faster and easier way to access information required to break full-disk encryption by booting from a flash drive, extracting the system’s hibernation keys and obtaining encryption metadata required to brute-force the original plain-text passwords to encrypted volumes.

The article provide following information for forensics experts:

• Dealing with full disk encryption
• No encryption: do i still need a password?
• Why not reset the password?
• Launching the chain reaction
• Recovering windows logon password

Whether or not you are able to recover the password with a preliminary attack is a matter of luck, especially if there are multiple users or a domain controller in place. If you were unable to recover the particular user’s Windows logon password, as a last resort, you may reset the account password (or remove account lock if it was locked).

Laura Heisman is an author of this article about how they have built and managed very successful global teams through the years and share some great insight on how to work with dispersed teams around the world.

Successfully managing a global team requires rock star leadership and strong communication skills, coupled with cultural awareness and empathy. It’s important to collaborate with your teams in ways that engage employees to work together as one team across multiple departments, languages, time zones, and cultures.

The article covers topics, like:

• What best practices have worked well for you?
• What is the impact of COVID-19 when it comes to remote working?
• How do you work across different time zones and continents?
• How do you work across different time zones and continents?
• Any advice for new remote workers in similar functions to yours?

Plan your home office and your work-week, and be sure to leverage tools and technology available to accomplish this. It was also recommended that you have a divide between your ‘home’ and ‘office’. Good read!

Railway Oriented Programming (ROP) is a functional programming technique that allows sequential execution of functions, not necessarily synchronous. The key concept is that each function can only accept and return Container of either Success or Failure. Failure wraps Throwable type and Success can be of any type. Published by Tonje Krosby.

This concept is present but not promoted in Kotlin since Kotlin 1.3. The matter of fact Kotlin KEEP states that feature may change in the future. At the time of writing this blog post, Result is a Container-like type which can be used inside generic containers such as Collection.

Example of functions that can have this paradigm and how to write adapters to those functions:

• Single track functions. Functions with no errors and no possible exceptions. See ComplexNumber and ComplexNumberResult examples.
• Dead-end functions. Functions which throw errors. Such as OutOfMemoryError or StackOverflowError. See HtmlParser and HtmlParserResult examples.
• Functions that throw exceptions. This is especially useful with IO operations. See DownloadPage examples.
• Supervisory functions. Functions which logs and supervise existing code, not necessarily return the result. See SentryLogin and SentryLoginResult examples.

Excellent read for all fans of functional programming!

In this post, we’ll look at the new loading attribute which brings native img and iframe lazy-loading to the web! Web pages often contain a large number of images, which contribute to data-usage, page-bloat and how fast a page can load. By Addy Osmani, Engineering Manager at Google working on Chrome.

Many of these images are offscreen, requiring a user to scroll in order to view them. Historically, to limit the impact offscreen images have on page load times, developers have needed to use a JavaScript library (like LazySizes) in order to defer fetching these images until a user scrolls near them.

What if the browser could avoid loading these offscreen images for you? This would help content in the view-port load quicker, reduce overall network data usage and on lower-end devices, reduce memory usage. Well, I’m happy to share that this will soon be possible with the new loading attribute for images and iframes.

The article then goes and describes the loading attribute. The loading attribute allows a browser to defer loading offscreen images and iframes until users scroll near them. For details, including demo video follow the link to the original article. Nice one!