Posts on Renato L. de F. Cunha

OpenID Connect Authentication with Booked Scheduler

Wed, 04 Sep 2019 20:04:33 +0000

In this post I'll describe the steps for implementing a plugin for using AppID authentication with Booked Scheduler deployed as a Cloud Foundry application at the IBM Cloud.

It looks like PHP simply isn’t going to die, especially because there are some fine PHP applications such as Booked Scheduler. It looks good and works really well, and that’s what we’ve been using to schedule meeting rooms at the Brazil Research Lab for the last 5 years. On top of that, booked has a plugin system that’s really easy to use. Also, notice that although we’re using IBM Cloud here, one can easily customize this to any other environment.

An argument in favor of strong scaling for deep neural networks with small datasets

Fri, 21 Sep 2018 18:42:24 +0000

In this post, I'm writing about my most recent paper, which is joint work with Eduardo R. Rodrigues, Matheus Palhares Viana, and Dario Augusto Borges Oliveira. If you prefer videos, there's a pretty comprehensive version below. You can read the full paper on the arXiv.

Apart from all the hype that is certainly involved in Machine Learning, Deep Learning, and Artificial Intelligence these days, there is a lot of very good research being done. Since much of this research comes from big cloud players such as Amazon, Google, IBM, and Microsoft, it is natural for such research to favor big clusters with humongous datasets. This is also good marketing: we all want to have access to the biggest machines possible, and it is good business for a cloud provider to know how to handle big machine learning workloads.

Non-stationarity and bandit algorithms

Thu, 22 Feb 2018 18:47:41 +0000

I’ve been (re)reading the second edition of Rich Sutton’s Introduction to Reinforcement Learning, and I’ve decided to do the programming exercises. (Previously, I followed Fermat’s tradition of solving exercises on the margin.)

Exercise 2.5 is nice. It asks us to demonstrate the difficulties that sample-average methods have for nonstationary problems by experimentation. The intuition being that, as current distribution means deviate from their original values, the bandit algorithm takes longer to update its estimates with regards to what it considers the optimal action. Moving-average methods which give more weight to more recent returns tend to perform better.

Rock, Paper, Starcraft review

Wed, 01 Mar 2017 19:05:40 +0000

Had forgotten to post this here. Enjoy! (YouTube video.)

Simple image classification using TensorFlow and CIFAR-10

Wed, 01 Mar 2017 19:03:44 +0000

Almost one year after following cs231n online and doing the assignments, I met the CIFAR-10 dataset again.

This time, instead of implementing my Convolutional Neural Network from scratch using numpy, I had to implement mine using TensorFlow, as part of one of the Deep Learning Nano Degree assignments.

As an aside, since this course uses some content from the free Deep Learning course, they took the time to fix the sub par presentation of that course. Good. :-)

Deploying JupyterHub behind an Apache reverse proxy’s subdirectory

Tue, 29 Nov 2016 19:34:32 +0000

JupyterHub supports deployment behind a reverse proxy and even instructs how to do so behindnginx in its manual. I was unable to find documentation about how to do this serving the content from a reverse proxy subdirectory.

Since there’s been a long time since I worked as a sysadmin and, therefore, I tend to forget these things, here hoes my documentation, in the hopes it will be useful to someone else.

Checking whether a pointer is valid in Linux

Wed, 09 Dec 2015 19:36:42 +0000

Recently I noticed some calls I was making to a certain C API were returning pointers that I thought were invalid. A quick inspection with gdb confirmed my fears. A particular interaction showed:

(gdb) print *job->someMember
Cannot access memory at address 0xec00000005

Unfortunately I don’t have access to the source of the API and won’t be able to change it. So now I’m at a loss. How can I know if a pointer is valid without dereferencing it? :S

Implementing a reader-writer lock for multiple processes in python

Fri, 27 Nov 2015 19:38:27 +0000

Python lacks a lock type that supports multiple readers and a single writer across multiple processes. Here we investigate how one can be implemented using only default python features: ctypes and mmap, and the POSIX reader-writer locks.

TL;DR: If you just want to use it, refer to the module’s repository.

Introduction Link to heading

I had a setting similar to the figure below. A single Python process would start up and spawn N children that would receive requests from the network. All these children shared a memory region from which they would read data for processing, with the parameters that defined how the processing should be done coming from the network.

A review of Udacity’s Machine Learning Engineer nanodegree

Thu, 19 Nov 2015 19:41:38 +0000

You can say I’m into AI. I’ve published a couple of papers on the subject and worked with Game AI before. Moreover, I’ve taken the original AI and ML classes offered by Stanford University back in 2011. Still, I felt my memory could use some refreshing, and decided to give the program a shot.

There are some reviews of other nanodegree programs on the Internet, so I’m going to talk about the Machine Learning one specifically.

How to connect your Android devices to an OpenVPN VPN

Sun, 28 Dec 2014 19:44:27 +0000

There are various OpenVPN configuration tutorials around the Internet, this post aims to fill in the gaps on how to configure the OpenVPN server, and OpenVPN for Android clients, while managing a simple firewall configured with UFW running an Arch Linux system.

Introduction Link to heading

OpenVPN is a robust and highly flexible VPN daemon. OpenVPN supports SSL/TLS security, Ethernet bridging, TCP or UDP tunnel transport through proxies or NAT, support for dynamic IP addresses and DHCP, scalability to hundreds or thousands of users, and portability to most major OS platforms. Further details about OpenVPN (and advanced tips on how to to configure it) can be found in the ArchWiki OpenVPN entry.

Mysterious 501 error responses and WebSEAL

Sun, 02 Mar 2014 19:46:37 +0000

This is just a quick post for future reference. Hopefully, it can help others in figuring out why a service behind Tivoli Access Manager's WebSEAL is returning a 501 error from POST requests.

Recently, I wrote a client using Apache's HttpClient for accessing a REST API protected by IBM's WebSEAL 6. While uploading data via POST, I started receiving a mysterious 501 Not Implemented error.

The data in question was being uploaded with chunked encoding, as in the following snippet¹.

Making the Playstation Eye’s microphones work on Linux

Thu, 26 Apr 2012 19:48:07 +0000

Introduction Link to heading

I’m the proud owner of a PlayStation® Move system and also a proud user of the Arch Linux distribution. As much as pride is concerned, it’s not very “proudful” when you decide to give your PlayStation® Eye (PS3 Eye) a try as a webcam, and you realize it doesn’t quite work as you expected. This post investigates this issue and proposes a solution.

Please, don’t use exit(-1)

Sat, 01 Oct 2011 19:50:05 +0000

Introduction Link to heading

In the past two weeks I saw a few cases where programmers were telling their programs to exit(), but without fully understanding what their exit codes meant and how they were represented. In this post I’ll explain what I believe to be the right way of doing it.

The first time I started to think about this was when I saw a script similar to the one below. But what really motivated my was [this reddit thread][]. If a practice I believe is bad starts to appear on best practices guides, then it’s time to express my opinion, right? [this reddit thread]: http://www.reddit.com/r/linux/comments/kwaoc/bash_best_practices/

Cognitive biases

Sun, 11 Jul 2010 19:51:45 +0000

Ever since I first heard about the Dunning-Kruger effect, I’ve started observing a bit more how I and some other people behave. Last week a friend introduced the “Impostor syndrome” concept to me, a “syndrome” apparently common among graduate students. There’s a very interesting article that presents the concept nicely, and which I quote here:

“Impostor syndrome” is the name given to the feelings that Abigail and many other young scientists describe: Their accomplishments are just luck or deceit, and they’re in over their heads. The key to getting past it, experts say, is making accurate, realistic assessments of your performance. Perhaps equally important: knowing you’re not alone. Abigail thinks that sharing her feelings with other people is how she will eventually come to grips with her sense of feeling like an impostor. “It’s fantastic to hear other people say, ‘I’ve felt that way, too.’ "

One month and a half in the Summer of Code program

Fri, 09 Jul 2010 19:55:10 +0000

It’s been a month and a half since I started working on the Mercurial project as part of the Google Summer of Code program. Mercurial is a free, distributed source control tool written in Python that efficiently handles projects of any size and offers an easy and intuitive interface while the Google Summer of Code is a global program that offers student developers stipends to write code for various open source software projects.