Did you get a flu shot this year? How about last year? In a world of next-day delivery and instant downloads, making the yearly pilgrimage to the doctor or the minute clinic feels like an outdated concept. Even if you get your shots free at the office, it’s still a pain to have to get vaccinated every year.
Unfortunately, there’s really no other way to deal with the annual threat of influenza. There’s no single vaccine for the flu because there are multiple strains that are always mutating. Unlike other viruses with one-and-done vaccinations, influenza is a moving target. Developing, producing, and distributing millions of vaccines every year is a massive operation that never stops, or even slows down a little bit. It’s basically Santa Claus territory — if Santa Claus delivered us all from mass epidemics.
The numbers are staggering. For the 2018-19 season, as in last year, there were 169.1 million doses distributed in the United States, up from 155.3 million doses the year before. How do they do it? We’re gonna roll up our sleeves and take a stab at it.
When exploring the realm of Machine Learning, it’s always nice to have some real and interesting data to work with. That’s where the bats come in – they’re fascinating animals that emit very particular ultrasonic calls that can be recorded and analysed with computer software to get a fairly good idea of what species they are. When analysed with an FFT spectogram, we can see the individual call shapes very clearly.
Nyctalus noctula (noctule bat)
Myotis nattereri (natterera’s bat)
Plecotus auritus (brown long eared)
Pipistrellus pipistrellus (common pipistrelle)
Creating an open source classifier for bats is also potentially useful for the world outside of Machine Learning as it could not only enable us to more easily monitor bats themselves, but also the knock on effects of modern farming methods on the natural environment. Bats feed on moths and other night flying insects which themselves have been decimated in numbers. Even in the depths of the countryside here in the UK these insects are a fraction of the population that they used to be 30 years ago, but nobody seems to have monitored this decline.
So getting back to our spectograms, it would be perfectly reasonable to throw these images at a convolutional neural network (CNN) and use an image feature-recognition strategy. But I wanted to explore the depths of the mysterious Random Forest. Continue reading “Training Bats In The Random Forest With The Confusion Matrix”→
Back in the days before kids could be placated with a $50 Android burner phone, many a youngster was gifted a so-called “educational computer” to keep them occupied. Invariably looking like a fever dream version of the real computer their parents didn’t want to let them use, these gadgets offered monochromatic exploits that would make Zork look like Fortnite. Due equally to their inherent hardware limitations and the premise of being an educational toy, the “games” on these computers often took the form of completing mathematical equations or answering history questions.
The VTech PreComputer 1000 is a perfect specimen of this particular style of educational toy. Released in 1988, it was advertised as a way for pre-teens to become more comfortable with operating a real computer; since at that point, it had become abundantly clear that the coming decade would see a beige box on every professional’s desk. Its full-size QWERTY keyboard was specifically mentioned in the product’s accompanying literature as a way to get young hands accustomed to the ways of touch typing.
Words of wisdom from the PreComputer 1000’s manual.
By the mid-1990s these devices would have progressed far enough to include passable text-to-speech capabilities and primitive graphics, but the junior professional who found him or herself seated in front of the PreComputer 1000 was treated to a far more spartan experience. It’s perhaps just as well that this particular educational computer was listed as a training tool, because even in 1988, surely a session with this toy must have felt very much like work.
But that’s not to say the PreComputer 1000 is without its own unique charms. In an effort to help cement its role as a “trainer” for more conventional computers, VTech saw fit to equip the PreComputer with its own BASIC interpreter. They even included generous written documentation that walked young programmers through the various commands and functions. Even today, there’s something oddly appealing about a mobile device with a full keyboard that can run BASIC programs for better than 24 hours on batteries (even if they’re alkaline “C” cells).
Let’s take a look inside this more than 30 year old mobile device, and see how the designers managed to create a reasonable facsimile of actual computing on a kid-friendly budget.
There’s a newish development in the world of keyboards; the optical switch. It’s been around for a couple years in desktop keyboards, and recently became available on a laptop keyboard as well. These are not replacements for your standard $7 keyboard with rubber membrane switches intended for puttering around on your raspberry pi. Their goal is the gamer market.
The question, though, is are these the equivalent of Monster Cables for audiophiles: overpriced status symbols? Betteridge would be proud; the short answer is that no, there is a legitimate advantage, and for certain types of use, it makes a lot of sense.
If you count yourself among the several hundred of our closest friends that have joined us at Supplyframe HQ for the 2019 Hackaday Superconference, then by now you’ll have your hands on one of this year’s incredible FPGA badges. It should come as no surprise that an incredible amount of time and effort went into developing and manufacturing this exceptionally unique piece of hardware; the slick gadget in your hands today took nearly an entire year to develop, and work continued on it until very literally the last possible moment.
Badge designer Jeroen Domburg (aka Sprite_TM), Hackaday staff, and a team of dedicated volunteers were still putting the final touches on these ambitious devices less than 24 hours before they were distributed to the first wave of Superconference attendees. Naturally, that’s not exactly how things were supposed to go. But when you’ve got a group of people that want to push the envelope and build something truly incredible, convincing them to actually stop working can be a challenge in itself.
In fact, development of the badge is still ongoing. Fixes and improvements are being made to the software even as you read this, and if you haven’t already, you should upgrade your badge to make sure you’ve got the latest and greatest from our international team of wizards. We all know that conference badges have an unfortunate habit of languishing on the shelf and collecting dust, but the 2019 Superconference badge was built to challenge you for longer than just one weekend. Consider yourself warned: for every Supercon badge that gets tossed in a drawer come Monday, Sprite_TM will shed a single tear.
After the break, come along as we turn back the clock and take a look at the last minute dash to get 500+ badges programmed and ready to go before the doors opened for the 2019 Hackaday Superconference.
If you came here from an internet search because your battery just blew up and you don’t know how to put out the fire, then use a regular fire extinguisher if it’s plugged in to an outlet, or a fire extinguisher or water if it is not plugged in. Get out if there is a lot of smoke. For everyone else, keep reading.
I recently developed a product that used three 18650 cells. This battery pack had its own overvoltage, undervoltage, and overcurrent protection circuitry. On top of that my design incorporated a PTC fuse, and on top of that I had a current sensing circuit monitored by the microcontroller that controlled the board. When it comes to Li-Ion batteries, you don’t want to mess around. They pack a lot of energy, and if something goes wrong, they can experience thermal runaway, which is another word for blowing up and spreading fire and toxic gasses all over. So how do you take care of them, and what do you do when things go poorly?
For all the successes of modern weather forecasting, where hurricanes, blizzards, and even notoriously unpredictable tornadoes are routinely detected before they strike, reliably predicting one aspect of nature’s fury has eluded us: earthquakes. The development of plate tectonic theory in the middle of the 20th century and the construction of a worldwide network of seismic sensors gave geologists the tools to understand how earthquakes happened, and even provided the tantalizing possibility of an accurate predictor of a coming quake. Such efforts had only limited success, though, and enough false alarms that most efforts to predict earthquakes were abandoned by the late 1990s or so.
It may turn out that scientists were looking in the wrong place for a reliable predictor of coming earthquakes. Some geologists and geophysicists have become convinced that instead of watching the twitches and spasms of the earth, the state of the skies above might be more fruitful. And they’re using the propagation of radio waves from both space and the ground to prove their point that the ionosphere does some interesting things before and after an earthquake strikes.
