Render Yourself Invisible To AI With This Adversarial Sweater Of Doom

October 20, 2022 by 57 Comments

Ugly sweater season is rapidly approaching, at least here in the Northern Hemisphere. We’ve always been a bit baffled by the tradition of paying top dollar for a loud, obnoxious sweater that gets worn to exactly one social event a year. We don’t judge, of course, but that’s not to say we wouldn’t look a little more favorably on someone’s fashion choice if it were more like this AI-defeating adversarial ugly sweater.

The idea behind this research from the University of Maryland is not, of course, to inform fashion trends, nor is it to create a practical invisibility cloak. It’s really to probe machine learning systems for vulnerabilities by making small changes to the input while watching for changes in the output. In this case, the ML system was a YOLO-based vision system which has little trouble finding humans in an arbitrary image. The adversarial pattern was generated by using a large set of training images, some of which contain the objects of interest — in this case, humans. Each time a human is detected, a random pattern is rendered over the image, and the data is reassessed to see how much the pattern lowers the object’s score. The adversarial pattern eventually improves to the point where it mostly prevents humans from being recognized. Much more detail is available in the research paper (PDF) if you want to dig into the guts of this.

The pattern, which looks a little like a bad impressionist painting of people buying pumpkins at a market and bears some resemblance to one we’ve seen before in similar work, is said to work better from different viewing angles. It also makes a spiffy pullover, especially if you’d rather blend in at that Christmas party.

Hackaday Wants You: Be A Supercon Volunteer

October 5, 2022 by 4 Comments
Spot the volunteers! (Hint: red shirts. And you know what happens to the red shirts…)

The Supercon approaches! If you are thinking of attending, but the cost of admission is too steep, one way to get in for free is to volunteer. That’s three wonderful days of events, two nights of partying, lunch, dinner, and of course Supercon. All you have to do is help us run the show.

Volunteers help out all around, giving out schwag bags, hustling speakers here and there, and just generally working behind the scenes to make Supercon super. We’re looking for three four-hour shifts over the whole long weekend,

So if you’re interested in helping out, and you’d like to get in free and get super volunteer-only gear to boot, put in your application now. We’ll be accepting volunteers until October 20th and getting in touch by email on October 24th.

Of course, we just announced the first round of speakers, we’ve got the badge reveal coming up, and much, much more. Follow along here, or at Hackaday.io/superconference for more info.

Lithium-Ion Batteries Are Easy To Find

October 3, 2022 by 43 Comments

In the first article, I’ve given you an overview of Lithium-Ion batteries and cells as building blocks for our projects, and described how hackers should treat their Lithium-Ion cells. But what if you don’t have any LiIon cells yet? Where do you get LiIon cells for your project?

Taking laptop batteries apart,  whether the regular 18650 or the modern pouch cell-based ones, remains a good avenue – many hackers take this road and the topic is extensively covered by a number of people. However, a 18650 cell might not fit your project size-wise, and thin batteries haven’t quite flooded the market yet. Let’s see what your options are beyond laptops. Continue reading “Lithium-Ion Batteries Are Easy To Find” →

2022 Hackaday Prize: Congratulations To The Winners Of The Climate-Resilient Communities Challenge

September 28, 2022 by 1 Comment

Holy humanitarian hacking, Batman! We asked you to come up with your best climate-forward ideas, and you knocked it out of the ionosphere! Once again, the judges had a hard time narrowing down the field to just ten winners, but they ultimately pulled it off — and here are the prize-winning projects without much further ado.

In the Climate-Resilient Challenge, we asked you to design devices that help build communities’ resilience to severe weather and the increasing frequency of natural disasters due to climate change, and/or devices that collect environmental data that serves as hard evidence in the fight for changes in local infrastructure. While several people focused on air quality, which is something we tend to think of as a human need, plenty others thought of the flora and fauna with which we share this planet.

Continue reading “2022 Hackaday Prize: Congratulations To The Winners Of The Climate-Resilient Communities Challenge” →

Neon Lamps — Not Just For Pilot Lights

September 13, 2022 by 30 Comments

It’s easy to see why LEDs largely won out over neon bulbs for pilot light applications. But for all the practical utility of LEDs, they’re found largely lacking in at least one regard over their older indicator cousins: charm. Where LEDs are cold and flat, the gentle orange glow of a neon lamp brings a lot to the aesthetics party, especially in retro builds.

But looks aren’t the only thing these tiny glow lamps have going for them, and [David Lovett] shows off some of the surprising alternate uses for neon lamps in his new video. He starts with an exploration of the venerable NE-2 bulb, which has been around forever, detailing some of its interesting electrical properties, like the difference between the voltage needed to start the neon discharge and the voltage needed to maintain it. He also shows off some cool neon lamp tricks, like using them for all sorts of multi-vibrator circuits without anything but a few resistors and capacitors added in. The real fun begins when he breaks out the MTX90 tube, which is essentially a cold cathode thyratron. The addition of a simple control grid makes for some interesting circuits, like single-tube multi-vibrators.

The upshot of all these experiments is pretty clear to anyone who’s been following [David]’s channel, which is chock full of non-conventional uses for vacuum tubes. His efforts to build a “hollow state” computer would be greatly aided by neon lamp circuits such as these — not to mention how cool they’d make everything look.

Continue reading “Neon Lamps — Not Just For Pilot Lights” →

Machine Learning App Remembers Names So You Don’t Have To

September 3, 2022 by 9 Comments

Depending on your point of view, real-life conversations with strangers can either be refreshing or terrifying. Some of us are glib and at ease in new social situations, while others are sure that the slightest flub will haunt them forever. And perhaps chief among these conversational faux pas is forgetting the name of the person who just introduced themselves a few seconds before.

Rather than commit himself to a jail of shame on such occasions, [Caleb] fought back with this only slightly creepy name-recalling smartphone app. The non-zero creep factor comes from the fact that, as [Caleb] points out, the app crosses lines that most of us would find unacceptable if Google or Amazon did it — like listening to your every conversation. It does this not to direct ads to you based on your conversations, but to fish out the name of your interlocutor from the natural flow of the conversation.

It turns out to be a tricky problem, even with the help of named-entity recognition (NER), which basically looks for the names of things in natural text. Apache OpenNLP, the NER library used here, works well at pulling out names, but figuring out whether they’re part of an introduction or just a bit of gossip about a third party is where [Caleb] put the bulk of his coding effort. That, and trying to make the whole thing at least a little privacy-respecting. See the video below for a demo.

To be sure, this doesn’t do much more than a simple, ‘remind me of your name again?’ would, but without the embarrassment. It’s still pretty cool though, and we’re especially jazzed to learn about NER and the tons of applications for it. Those are projects for a future day, though. We’re just glad to see that [Caleb] has moved on from monitoring the bodily functions of his dog and his kid. At least for now.

Continue reading “Machine Learning App Remembers Names So You Don’t Have To” →

LoRa Air Quality Monitor Raises The Bar On DIY IoT

August 30, 2022 by 15 Comments

We’ve seen an incredible number of homebrew environmental monitors here at Hackaday, and on the whole, they tend to follow a pretty predicable pattern. An ESP8266 gets paired with a common temperature and humidity sensor, perhaps a custom PCB gets invited to the party, and the end result are some values getting pushed out via MQTT. It’s a great weekend project to get your feet wet, but not exactly groundbreaking in 2022.

Which is why we find the AERQ project from [Mircea-Iuliu Micle] so refreshing. Not only does this gadget pick up temperature and humidity as you’d expect, but its Bosch BME688 sensor can also sniff out volatile organic compounds (VOCs) and gases such as carbon monoxide and hydrogen. The datasheet actually claims this is the “first gas sensor with Artificial Intelligence (AI)”, and while we’re not sure what exactly that means in this context, it’s a claim that apparently warrants a price tag of $15+ USD a pop in single quantities.

There’s an AI hiding in there someplace.

But the fancy sensor isn’t the only thing that sets AERQ apart from the competition. Instead of a member of the ubiquitous ESP family, it’s using the Wio-E5, a relatively exotic STM32 package that integrates a long-range LoRa radio. [Mircea-Iuliu] has paired that with a Linx USP-410 chip antenna or, depending on which version of the four-layer PCB you want to use, a u.Fl connector for an external antenna. The whole thing is powered by a simple USB connection, and its Mbed OS firmware is setup to dump all of its collected data onto The Things Network.

All told, it’s a very professional build that certainly wouldn’t look out of place if it was nestled into some off-the-shelf air quality monitor. While the high-end detection capabilities might be a bit overkill for home use, [Mircea-Iuliu Micle] points out that AERQ might provide useful insight for those running indoor events as COVID-19 transitions into its endemic stage.