Menu
"hardware" entries
Graph Algorithm, Touchy Robots, Python Bolt-Ons, and Building Data Products
-
Exact Maximum Clique for Large or Massive Real Graphs — explanation of how BBMCSP works.
-
Giving Robots and Prostheses the Human Touch — the team, led by mechanical engineer Veronica J. Santos, is constructing a language of touch that both a computer and a human can understand. The researchers are quantifying this with mechanical touch sensors that interact with objects of various shapes, sizes, and textures. Using an array of instrumentation, Santos’ team is able to translate that interaction into data a computer can understand. The data is used to create a formula or algorithm that gives the computer the ability to identify patterns among the items it has in its library of experiences and something it has never felt before. This research will help the team develop artificial haptic intelligence, which is, essentially, giving robots, as well as prostheses, the “human touch.”
-
boltons — things in Python that should have been builtins.
-
Everything We Wish We’d Known About Building Data Products (DJ Patil and RusJan Belkin) — Data is super messy, and data cleanup will always be literally 80% of the work. In other words, data is the problem. […] “If you’re not thinking about how to keep your data clean from the very beginning, you’re fucked. I guarantee it.” […] “Every single company I’ve worked at and talked to has the same problem without a single exception so far — poor data quality, especially tracking data,” he says.“Either there’s incomplete data, missing tracking data, duplicative tracking data.” To solve this problem, you must invest a ton of time and energy monitoring data quality. You need to monitor and alert as carefully as you monitor site SLAs. You need to treat data quality bugs as more than a first priority. Don’t be afraid to fail a deploy if you detect data quality issues.
Augmenting Humans, Body-Powered CPUs, Predicting the Future, and Hermit Life
-
Unpowered Ankle Exoskeleton — “As we understand human biomechanics better, we’ve begun to see wearable robotic devices that can restore or enhance human motor performance,” says Collins. “This bodes well for a future with devices that are lightweight, energy-efficient, and relatively inexpensive, yet enhance human mobility.”
-
Body-Powered Processing (Ars Technica) — The new SAM L21 32-bit ARM family of microcontroller (MCUs) consume less than 35 microamps of power per megahertz of processing speed while active, and less than 200 nanoamps of power overall when in deep sleep mode—with varying states in between. The chip is so low power that it can be powered off energy capture from the body. (via Greg Linden)
-
Temporal Effects in Trend Prediction: Identifying the Most Popular Nodes in the Future (PLOSone) — We find that TBP have high general accuracy in predicting the future most popular nodes. More importantly, it can identify many potential objects with low popularity in the past but high popularity in the future.
-
The Shut-In Economy — In 1998, Carnegie Mellon researchers warned that the Internet could make us into hermits. They released a study monitoring the social behavior of 169 people making their first forays online. The Web-surfers started talking less with family and friends, and grew more isolated and depressed. “We were surprised to find that what is a social technology has such anti-social consequences,” said one of the researchers at the time. “And these are the same people who, when asked, describe the Internet as a positive thing.”
Tuning Fanout, Moore's Law, 3D Everything, and Social Graph Analysis
-
Facebook’s Mystery Machine — The goal of this paper is very similar to that of Google Dapper[…]. Both work [to] try to figure out bottlenecks in performance in high fanout large-scale Internet services. Both work us[ing] similar methods, however this work (the mystery machine) tries to accomplish the task relying on less instrumentation than Google Dapper. The novelty of the mystery machine work is that it tries to infer the component call graph implicitly via mining the logs, where as Google Dapper instrumented each call in a meticulous manner and explicitly obtained the entire call graph.
-
The Multiple Lives of Moore’s Law — A shrinking transistor not only allowed more components to be crammed onto an integrated circuit but also made those transistors faster and less power hungry. This single factor has been responsible for much of the staying power of Moore’s Law, and it’s lasted through two very different incarnations. In the early days, a phase I call Moore’s Law 1.0, progress came by “scaling up”—adding more components to a chip. At first, the goal was simply to gobble up the discrete components of existing applications and put them in one reliable and inexpensive package. As a result, chips got bigger and more complex. The microprocessor, which emerged in the early 1970s, exemplifies this phase. But over the last few decades, progress in the semiconductor industry became dominated by Moore’s Law 2.0. This era is all about “scaling down,” driving down the size and cost of transistors even if the number of transistors per chip does not go up.
-
BoXZY Rapid-Change FabLab: Mill, Laser Engraver, 3D Printer (Kickstarter) — project that promises you the ability to swap out heads to get different behaviour from the “move something in 3 dimensions” infrastructure in the box.
-
SociaLite (Github) — a distributed query language for graph analysis and data mining. (via Ben Lorica)
Welfare and Entrepreneurialism, Infrastructure Secrets, Insectoid Robots, Hacking Hexbugs
-
Welfare Makes America More Entrepreneurial (The Atlantic) — In a 2014 paper, Olds examined the link between entrepreneurship and food stamps, and found that the expansion of the program in some states in the early 2000s increased the chance that newly eligible households would own an incorporated business by 16%. (Incorporated firms are a better proxy for job-creating startups than unincorporated ones.)
-
Security of Infrastructure Secrets — everything has a key that’s just one compromise or accidental drop away.
-
Festo’s Fantastical Insectoid Robots Include Bionic Ants and Butterflies (IEEE) — Each butterfly has a 50-centimeter wingspan and weighs just 32 grams, but carries along two servo motors to independently actuate the wings, an IMU, accelerometer, gyro, and compass, along with two tiny 90-mAh lithium-polymer batteries. With a wing beat frequency of between one and two flaps per second, top speed is 2.5 m/s, with a flight time of three to four minutes before needing a 15-minute charge. The wings themselves use impossibly thin carbon rods for structure, and are covered with an even thinner elastic capacitor film.
-
Arduino Celebration and Hexbugs hacking with Bob Martin (SparkFun) — The Hunter demo is a combination of object detection and object avoidance. It uses an IR sensor array to determine objects around it. Objects that appear and then disappear quickly, say in a second or two are targets which it will walk towards; however, a target that stays constant will be avoided. I’m still trying to find the perfect balance between making a decision between fleeing prey and a wall using only simple proximity samples from an IR detector array.
GPU Graph Algorithms, Data Sharing, Build Like Google, and Distributed Systems Theory
-
gunrock — a CUDA library for graph primitives that refactors, integrates, and generalizes best-of-class GPU implementations of breadth-first search, connected components, and betweenness centrality into a unified code base useful for future development of high-performance GPU graph primitives. (via Ben Lorica)
-
How to Share Data with a Statistician — some instruction on the best way to share data to avoid the most common pitfalls and sources of delay in the transition from data collection to data analysis.
-
Bazel — a build tool, i.e. a tool that will run compilers and tests to assemble your software, similar to Make, Ant, Gradle, Buck, Pants, and Maven. Google’s build tool, to be precise.
-
You Can’t Have Exactly-Once Delivery — not about the worst post office ever. FLP and the Two Generals Problem are not design complexities, they are impossibility results.
Tricorder Prototype, Web Performance, 3D Licensing, and Network Simulation
-
Tricorder Prototype — collar+earpiece, base station, diagnostic stick (lab tests for diabetes, pneumonia, tb, etc), and scanning wand (examine lesions, otoscope for ears, even spirometer). (via Slashdot)
-
Souders Joins Speedcurve — During these engagements, I’ve seen that many of these companies don’t have the necessary tools to help them identify how performance is impacting (hurting) the user experience on their websites. There is even less information about ways to improve performance. The standard performance metric is page load time, but there’s often no correlation between page load time and the user’s experience. We need to shift from network-based metrics to user experience metrics that focus on rendering and when content becomes available. That’s exactly what Mark is doing at SpeedCurve, and why I’m excited to join him.
-
3 Steps for Licensing Your 3D-Printed Stuff (PDF) — this paper is not actually about choosing the right license for your 3D printable stuff (sorry about that). Instead, this paper aims to flesh out a copyright analysis for both physical objects and for the digital files that represent them, allowing you to really understand what parts of your 3D object you are—and are not—licensing. Understanding what you are licensing is key to choosing the right license. Simply put, this is because you cannot license what you do not legally control in the first place. There is no point in considering licenses that ultimately do not have the power to address whatever behavior you’re aiming to control. However, once you understand what it is you want to license, choosing the license itself is fairly straightforward. (via BoingBoing)
-
Augmented Traffic Control — Facebook’s tool for simulating degraded network conditions.
Changing Behaviour, Building Filters, Public Access, and Working Capital
-
Using Monitoring Dashboards to Change Behaviour — [After years of neglect] One day we wrote some brittle Ruby scripts that polled various services. They collated the metrics into a simple database and we automated some email reports and built a dashboard showing key service metrics. We pinpointed issues that we wanted to show people. Things like the login times, how long it would take to search for certain keywords in the app, and how many users were actually using the service, along with costs and other interesting facts. We sent out the link to the dashboard at 9am on Monday morning, before the weekly management call. Within 2 weeks most problems were addressed. It is very difficult to combat data, especially when it is laid out in an easy to understand way.
-
Quiet Mitsubishi Cars — noise-cancelling on phone calls by using machine learning to build the filters.
-
NSF Requiring Public Access — NSF will require that articles in peer-reviewed scholarly journals and papers in juried conference proceedings or transactions be deposited in a public access compliant repository and be available for download, reading, and analysis within one year of publication.
-
Filtered for Capital (Matt Webb) — It’s important to get a credit line [for hardware startups] because growing organically isn’t possible — even if half your sell-in price is margin, you can only afford to grow your batch size at 50% per cycle… and whether it’s credit or re-investing the margin, all that growth incurs risk, because the items aren’t pre-sold. There are double binds all over the place here.
Billion Node Graphs, Asynchronous Systems, Deep Learning Hardware, and Vision Resources
-
Mining Billion Node Graphs: Patterns and Scalable Algorithms (PDF) — slides from a CMU academic’s talk at C-BIG 2012.
-
There Is No Now — One of the most important results in the theory of distributed systems is an impossibility result, showing one of the limits of the ability to build systems that work in a world where things can fail. This is generally referred to as the FLP result, named for its authors, Fischer, Lynch, and Paterson. Their work, which won the 2001 Dijkstra Prize for the most influential paper in distributed computing, showed conclusively that some computational problems that are achievable in a “synchronous” model in which hosts have identical or shared clocks are impossible under a weaker, asynchronous system model.
-
Deep Learning Hardware Guide — One of the worst things you can do when building a deep learning system is to waste money on hardware that is unnecessary. Here I will guide you step by step through the hardware you will need for a cheap high performance system.
-
Awesome Computer Vision — curated list of computer vision resources.
Working Manager, Open Source Server Chassis, Data Context, and Coevolved Design & Users
-
As a Working Manager (Ian Bicking) — I look forward to every new entry in Ian’s diary, and this one didn’t disappoint. But I’m a working manager. Is now the right time to investigate that odd log message I’m seeing, or to think about who I should talk to about product opportunities? There’s no metric to compare the priority of two tasks that are so far apart. If I am going to find time to do development I am a bit worried I have two options: (1) Keep doing programming after hours; (2) Start dropping some balls as a manager.
-
Introducing Yosemite (Facebook) — a modular chassis that contains high-powered system-on-a-chip (SoC) processor cards.
-
The Joyless World of Data-Driven Startups — There is so much invisible, fluid context wrapped around a data point that we are usually unable to fully comprehend exactly what that data represents or means. We often think we know, but we rarely do. But we really WANT it to mean something, because using data in our work is scientific. It’s not our decision that was wrong — we used the data that was available. Data is the ultimate scapegoat.
-
History of the Urban Dashboard — the dashboard and its user had to evolve in response to one another. The increasing complexity of the flight dashboard necessitated advanced training for pilots — particularly through new flight simulators — and new research on cockpit design.
Robot Swarms, Media Hacking, Inside-Out Databases, and Quantified Medical Self
-
Surgical Micro-Robot Swarms — A swarm of medical microrobots. Start with cm sized robots. These already exist in the form of pillbots and I reference the work of Paolo Dario’s lab in this direction. Then get 10 times smaller to mm sized robots. Here we’re at the limit of making robots with conventional mechatronics. The almost successful I-SWARM project prototyped remarkable robots measuring 4 x 4 x 3mm. But now shrink by another 3 orders of magnitude to microbots, measured in micrometers. This is how small robots would have to be in order to swim through and access (most of) the vascular system. Here we are far beyond conventional materials and electronics, but amazingly work is going on to control bacteria. In the example I give from the lab of Sylvain Martel, swarms of magnetotactic bacteria are steered by an external magnetic field and, interestingly, tracked in an MRI scanner.
-
Media Hacking — interesting discussion of the techniques used to spread disinformation through social media, often using bots to surface/promote a message.
-
Turning the Database Inside Out with Apache Samza — replication, secondary indexing, caching, and materialized views as a way of getting into distributed stream processing.
-
Apple Research Kit — Apple positioning their mobile personal biodata tools with medical legitimacy, presumably as a way to distance themselves from the stereotypical quantified selfer. I’m reminded of the gym chain owner who told me, about the Nike+, “yeah, maybe 5% of my clients will want this. The rest go to the gym so they can eat and drink what they want.”