HiI'm Roman from Porini Fooundation in Switzerland working mainly on blockchain for nature
Hi everyone, I'm a Senior GIS Analyst with the Royal Society for the Protection of Birds (Birdlife partner in the UK).
Hi y'all, Rob Appleby here...excited for this one!
Hi Rob. Nice to virtually meet you :)
You too Akiba!
Hi everyone. Thanks for joining.
Hi everyone, Liu Song from WWF China, working on Technology and Environment.
Hi I am a producer of Prowild wildlife detection systems
Hey, I'm Leslie Birch from Philadelphia-US. Working on Master's in Conservation Tech and love Arduino! :)
Hi all! I'm Hannah, Field Conservation Assistant at Chester Zoo, UK
Hello! I’m Caitlin, a postdoc at the University of Zurich, researching elephant behavior using survey images among other things.
Hello, I am Nobesuthu Adelaide Ngwenya, Ecologist - Zimbabwe Parks
Hi. I'm Akiba from Hackerfarm and FreakLabs. Everyone sounds like they're doing such interesting things :)
Hi, Vickie DeNicola with White Buffalo … we do applied research and management projects.
Ben Vander Velden - Denver, CO - no professional link at the moment, just passionate about technology and conservation.
Tanya Berger-Wolf from University of Illinois at Chicago and Wildbook @ Wild Me is here. Looking forward to everyone’s presentations and the followup discussion
Hi, I'm Laure Joanny, I'm a doctoral researcher currently working on the developement and impact of monitoring technologies for conservation law enforcement
Hi Everyone, Vance Russell from Ecosulis
Hej Hej everyone, Niklas from Northern Germany here. I am Technology advocate and studying Earth Science. Besides I am blogging about conservation technology: https://blog.niklasjordan.com/conservation-technology/
Alexandra Dumitrescu || WWF Romania
I’m Alexandra - coordinating the team for funding opportunities for Panda Labs - an innovation open experiment built on top of the WWF network
Hi all, I am a postdoc at Loughborough University, UK, looking at applying machine learning algorithms for visual recognition. I am interested in wildlife conservation.
Cheers, Mate Zec from Biom (BirdLife Croatia). We do a lot of GPS tracking of bird movement, as well as sound recording and drones for monitoring the extent illegal killing of birds
Hi Everyone I am Laurent Andriantsiferana, IOT Architect at CIsco, supporting Animal Conservation projects
Really great to meet everyone! Fantastic to hear all your different backgrounds :)
hi! i’m sarah. i work as a software engineer for an ecommerce company in south africa. looking to get into opensource conservation tech
More info about mataki here: http://mataki.org/
Great thread from Yvan about lesson learnt from mataki lite here: https://www.wildlabs.net/community/thread/573
and here: https://www.wildlabs.net/community/thread/574
The questions al just mentioned are what we summarised from your registrations to this meet up: https://www.wildlabs.net/community/thread/698
A recent update about the sea turtle tag: https://www.wildlabs.net/resources/case-studies/update-arribadas-low-cost-open-source-sea-turtle-tag
Hi - this is Nick Salafsky with Miradi Software. Interested in underlying creating data standards to enable info exchange to support conservation work....
Hi all! I am Pauline - R&D engineer at Sigfox Foundation. We have developed a rhino tracker to allow rangers to monitor and protect them to avoid poaching
Was Alasdair at GOSH in China?
Great presentation Alasdair
This is Katherine Zacarian from Planetary Health Solutions. We are developing implantable tracking tags for Great Apes. Thank you everyone. email@example.com.
They are open source hardware for scientific equipment
Ah! Thanks akiba
OK TY Alasdair :)
hi I'm aurèlie and I am trying to launch PandaSat! a satellite constellation for nature tracking, with support from Wildlabs tech hub!
many many questions for Alasdair a little later!
Arg you were quicker than me! Thanks :)
oh sorry. will hand the chat links to you :)
ha, not at all, please continue :D
I would love some way to connect with everyone on this call. First time I'm with people exactly in my interest field.
Definitely - I’d recommend joining this thread: https://www.wildlabs.net/community/thread/698
but we can discuss that in the bigger group - how to connect/collaborate
I think there are a lot of APIs that are available. I think open source by definition is source code availalbe and reproducible
Definitely agree. A good breakdown of what a platform needs to achieve.
Open source is more than just open an available source code. It means well engineered code base to which others can contribute. It means a well supported code base, versioning and testing.
I agree. I think open API is maybe developer friendly, but not open source.
Thanks Alasdair and David, fascinating discussion so far.
Here is the opportunity for the Trident grant program David mentioned: https://openexplorer.nationalgeographic.com/see
It's mainly semantics though. I've worked in both realms and I think everyone/company has their own path
are you open source but also a for-profit company? Or non-profit?
We are a non-profit and open source. Today, even big tech giants like Google and Microsoft are going open source on many of their solutions
How does everyone feel about the tension David menitoned re: profit and conservation?
> Today, even big tech giants like Google and Microsoft are going open source on many of their solutions
Google & Microsoft are in fact the biggest open source contributors on GitHub.com, at the moment!
tech companies benefit enormously from open-source...
I think conservation tech is poorly funded and many of the organizations that need to use tech have no funding for it and the foundations prefer to give money for actual field work rather than supporting conservation tech
Arribada is a C.I.C - Community Interest Company which is a not for profit (UK) that legally reinvests funds to support the development of open conservation tech.
C.I.C is a great model
If you’re in the US, a B Corp is a good alternative: https://bcorporation.net/
I think profit is very important for open source projects because financial sustainability is needed for longevity. But i think there are multiple paths to strike a balance
B Corps get an A+ from me!
Thanks Tanya. That's an important part. I think the tech needs to be a part of the conservation work. The funding should go to the conservation work, but the tech is part of it
We’ve had some success with funding from Conservation X Labs for early seed funding for R&D: https://conservationx.com/
Which was your project, Katie?
The funding for field work vs tech work reminds me a lot of the Roads and Bridges report, which was a report on sustainability of “digital infrastructure”, i.e. the open source projects that run the world. It makes the point that digital infrastructure is often invisible and doesn’t attract as much funding: https://www.fordfoundation.org/about/library/reports-and-studies/roads-and-bridges-the-unseen-labor-behind-our-digital-infrastructure/
That’s a great report - recommend it too.
Cool thanks for sharing Joe
Gonna drop this here just in case it's needed:https://www.wildbook.org
We are working on species recognition software and are starting a pilot within the coming months. perhaps thats a solution to process all that visual data
Gert- what application are you looking at? Any species/sort of data in particular?
We are going to implement it in our wildlife detection system especcialy designd for otter and deer
we can also use it for fish
Ah! Anywhere we can find out more?
definitely interested in ungulate solutions …
its brandnew, the brochures still need to be made :D
Tanya just mentioned rolling it out to elephants - a need more than one community member has flagged: https://www.wildlabs.net/community/thread/618
At Ol Pejeta,through the Conservation Technology Lab, we have have just installed some WPS covert cameras which are able to identify species, vehicles etc.
Ha! Far enough - please do share on WILDLABS when you have more info available
the technologies powering Wildbook is so fantastic. Have you considered using it to verify instances of illegal wildlife trafficking online?
Good question - I’ll ask Tanya
Okay I will
I work with camera trap networks across Africa and our team has implements some machine learning to use in conjunction with citizen science - works well for ungulates!
Is that through zooniverse Meredith - you’re with snapshot serengeti aren’t you?
Yes! I'm also working with WildCam Gorongosa and a number of smaller sites across South Africa. We've got a developer who has been integrating AI into our classification systems.
how do nature photographers and tourists engage with Wildbook to contribute their media to the braintrust? For example, have you reached out to Instagram — (I have contacts if you’d like)
Here's the GitHub: https://github.com/marco-willi/camera-trap-classifier
I will check that also meredith tx
Chimpface is one example of using machine learning to look at wildlife trafficking online/ via social media: https://www.wildlabs.net/resources/case-studies/chimpface-facial-recognition-combat-wildlife-trafficking
Tanya, do you have error rates or detection probabilities associated with your counts, or are they considered a census?
Will be there an recording of this meetup and will the slides be published?
Niklas - yes! They will be shared within a week
The recording, not sldies
But we will also include notes on presentations and main questions/ take-aways
Wow, that's fascinating. Never thought images would be used by poachers as well as scientists.
Tanya, in your opinion, is it feasible to create algorithms to extract even more specific data from the images for each individual? For example, the size of individuals or other traits such as age?
That was a great talk. Thank you, Tanya.
That was amazing, thanks Tanya.
@Tanya: Are the WildBook algorithms and models + coefficients open source?
Literally the question I was going to ask, pascal
Azure is the cloud service by Microsoft.
Can Wildbook also be used to identify wildlife from top-down aerial imagery (as opposed to ground-based?)
@Caitlin: There are currently algorithms that are estimating age for humans, so with a good enough quality image, in theory, it should be possible to do the same with animals that look significantly different as they age.
Have you come across Wildlife Insights? Is there an appreciable difference between Wildlife Insights and Wildbook?
Thomas - excellent question.
any concerns about deep fakes (e.g. poachers spoofing images of individuals to make it look there are more than there are or specific individuals are still alive when they aren't)?
Wildlife Insights is focusing on camera trap images and processing of data (inc video)
Right, that makes sense, thank you; we're planning some drone-based river dolphin surveys in North India later this year and being able to ID individuals from above would be ideal.
Do the animals need to have unique marks (e.g., tail notches, spots) or can facial structure be used ?
Thanks @Joannna, I have a paper in press showing that age can be estimated based on morphological features in elephants. Would be wonderful to automate it and scale it up!
Thank you Tanya for the great presentation. Can researchers have separate accounts on wildbook to manage their own data and research?
Oh shashank - you should talk to Marcello. He’s doing drone based river dolphin surveys in brazil
@Tanya: Does "Open source" also include the pre-trained models?
Tanya, with trained model, can we consider deploying this in distributed low-power device with camera ? something like rpi+gpu ?
I totally sympathize. Preparing a release takes a huge amount of time and effort.
From my perspective the limiting factor to open sourcing our hardware is the time resource required to write up/provide the tech support etc. Also its very much about paper output and to get things peer reviewed then the 'tech' often goes obsolete/becomes outdated (3yrs?)before publishing. This happened with Mataki. Should Wildlabs be the place to get stuff out as there seem to be many options??
+ 1 nigel!
I believe the model architectures are detailed in some published papers, so in theory it should be reproducible. See https://oar.princeton.edu/jspui/bitstream/88435/pr13m5n/1/Animal_Detection_Pipeline_for_Identification_2018.pdf
I would say WIldlabs + a sustainable income (various methods) to fund the core team that isn’t reliant on the core team having to run the support so they can focus on the product
@Stephanie: Yes, totally!
@Tanya: And please note that I am fully aware that it is much easier to ask someone else to open their source than it is to actually do it yourself...
So WILDLABS comprising ~2000-3000 people right? That's a pretty big company-equiv...I see something like a NASA model, having core teams working on projects with very wide support
I think a big component of it is the fragmentation between the different communities. Having a central place to communicate is critical.
Could not agree more!
I agree on a centralized community.
Yes, central place needed.
@Tanya, @Rob: Isn't that what the typical GitHub model is about? No project is forced to accept any particular contribution, but everybody is free to contribute and improve the project by creating their own synchronised copy ("fork").
A central place would be great to avoid replication but I think there is the problem of the current funding model, as mostly the wheel is reinvented due to people having to fundraise for themselves and the short timeframes. What is the best way to avoid that?
@nobesuthu: Yes, researchers can have separate accounts that will access only to their data. In fact, in Flukebook, charged by NOAA, we are implementing pretty sophisticated data sharing agreements among researchers
Conversation on open source projects here: https://www.wildlabs.net/community/thread/679
Hi Pascal, I think the problem with GitHub, is the project has already been done and resources poured into before realising 100 other people have similar repos...
plus it's really passive
If you can't maintain on Github, at least having a link for your project and description would be good. Lots of people check Github first before starting projects.
You definitely need to have strong boundaries and the ability to say no if you lead an open source project
Is this the one Joe? Mike McQuaid’s blog: https://mikemcquaid.com/
I’d recommend starting with that post
Rob: Sorry, not sure I get you: It seems to me that not putting a project on GitHub right at the start, the danger you mention only increases...
This is also really great for those of you looking to start/grow your project: https://mikemcquaid.com/2018/08/14/the-open-source-contributor-funnel-why-people-dont-contribute-to-your-open-source-project/
The title of the meetup was "low cost, open source solutions" but it doesn't seem low cost any more!
oops. @Rob, “plus it's really passive” what do you mean by passive?
Absolutely agree that organising / maintaining open source projects is hugely important and can (and often does) take a lot of effort. I think this is where public funding could come in.
Maybe I think of GitHub a bit like R (stats env), where generally projects are most of the way developed before it's posted, although obviously not always. I can see you point though.
I think for open source hardware, it's important to understand the manufacturing side of things
I agree Akiba!
I also think its pssible to have a business model that works for open source hardware. Typical margins at open source companies like Adafruit and Sparkfun are devices and parts are sold at around 3-5x BOM cost. The devices are still reasonably priced and companies are able to survive, matain, and support their efforts
@Tanya: I agree that there are lots of things to consider, particularly around sharing sensitive data. But I do think that SAAS is definitely _not_ open source.
Yeah Sparkfun and Ada are great examples of getting it to work...
Actually I helped LadyAda set up her first assembly line :)
@Pascal: I think there are levels of support. We open our code and if you don’t need our time to take it - great! But if you need engineering time to use our code then we think it’s fair to pay for that service
Akiba--I freelanced doing projects for Adafruit. Still a huge fan! :D
@Tanya: That sounds like an added-value-model - which is great to have in addition!
I feel like researchers and techies speak different languages figuratively. I think there needs to be education, workshops, videos that can bridge both groups.
@Akiba I think so too.
We've got philanthropic funding in India by demonstrating conservation impact not by developing new tech but by using off-the-shelf tech appropriately.
Can we have a space on Wildlabs that is about discussing new ideas for collaborating for funding? Where people can feedback on potential ideas?
These virtual meetups are a good starting point to bring everyone together :)
@Akiba - GOSH (Gathering of Open Science Hardware) helps to do that bridging.
I’m so on board with that idea @Rachael - will follow up.
@Shashank: In agriculture, we discuss the concept of "appropriate technology" a lot. A lot of groups try to throw IoT and drones to solve farming problems where the real work is in the distribution and logistics of the food
At the very first Open Hardware Workshop at CERN in 2011, there were companies saying that they were still very successful as a company although they fully shared all their hard- and software for niche scientific applications. Part of the argument was that if anyone is really interested in copying you, they will do so anyway - the difference would be just a few weeks at max. On the plus side, their expertise on their products would be so visible and evident that many customers would prefer to deal with them directly even more, even though there were cheaper knock-offs.
@shashank. We are now working on a bat tracking device using adapted acoustic devices. We are now able to make flight routes of bats visual. So you are right to involve producers in specific projects/products
@akiba yes exactly; in India half the job is telling people that they don't need to use drones!
@gert that sounds great, where could I find out more about your work?
I absolutely agree that open sourcing a project _should_ absolutely not entail a commitment to provide support for it, however desirable that might be from someone else's point of view.
@ shashank This project is also still ongoing and brandnew. We can keep in touch and you can find my contact info at www.prowild.net
@Thomas Starnes: yes, we are talking to Wildlife Insights. We were at the inaugural meeting
It's very easy to fall into the technology trap, while ignoring the larger picture. Conservation is inherently linked to communities. Finding solutions that integrate into their skills and values is very important for successful outcomes.
Who is that - working on bat tracking?
Gert I think?
Pascal, we are interested in various versions of available tech., from DIY from source code etc., then PCBs, to kits, to custom products. It's a time vs money equation for many people...
@Gert where can we read more about bat tracking?
Stepanie: Thanks - is it somewhere on the web?
@Laurent Andriantsiferana: yes, we are talking to Intel and Amazon about it. It’s a cutting edge research question of how to do it right
@Gert's info was posted up higher
still in development Htarold. We expect to do a trial later this year
Rob: Sounds great. I would love to see this space thrive!
A map of all the players and the tech they're working on would also be super helpful for roadmapping and collaboration
I agree Pascal. And there's space for all forms of systems
For those who are interested you can leave your contact info on my website www.prowild.net so I can keep ypou posted on the progress
That's a great idea Rachel
Gert - I recommend starting a thread on wild labs as well, so people can keep up to date
Thanks so much to our speakers and to everyone who showed up with great questions and enthusiasm!
Recording and notes will be posted here: https://www.wildlabs.net/virtual-meetups-season-2
I hope we will find a good solution for resolving the obvious tensions between conservation interests, private and scientific interests, and possibly even commercial interests!
Jump over to this thread to continue the conversation now: https://www.wildlabs.net/community/thread/698
thanks so much everyone...kicking A!
Thanks to WILDLABS and the speakers. Excellent webinar and chat!
thats also a good idea stephanie
Yes, thanks so much everyone!
Thanks speakers and organisers! Exciting stuff :)
Extremely inspiring stuff
Fantastic webinar and discussion, thanks everyone.
Thanks everyone, and thank you for organising this Steph! Really enjoyed this:)
Yes, it was great to hear and also virtually see everyone live. Thanks everyone!
Well done!!! Thank you
thanks Talia and Steph