Archive | Uncategorized RSS for this section

Training the Machines Update

Thanks for the great work helping us find fruits, flowers and unfolded leaves on Acer (the maples). That is going gangbusters and we hope you can continue the great pace of effort!

We wanted to pass along some good news about how well you did on the first training the machines expedition which focused on Prunus (cherries and allies). The results below are organized by “trait” and includes the total classifications, correct classifications and percentages. We also provide the true positives and true negatives as well and use that calculate what is called “sensitivity” and “specificity.” Sensitivity is a measure of the correct positive against the total true positives. And specificity is the same idea for true negatives. Accuracy is the overall ratio of correct classifications to total classifications from the gold standard. All of our assessments treat a hand-coded dataset that we created at Notes from Nature as a gold standard for comparison.
13691529605_073b9cf02d_k
Whenever we do these comparisons, we go back to our hand-coded dataset and check our hand-coded results against what appear to be mistakes. As it turns out, we had a number of mistakes in our “gold standard” that you helped us find. That is, the numbers are even better than what is reported below. You guys helped us find problems in our original “gold standard” scoring! That’s super important when developing supervised machine learning datasets. Overall, these are extremely encouraging results! We will be providing some further details about this work in follow up post but, again, this is very exciting news!
Unfolded leaves:

Overall:
Total classifications: 2998
Correct classifications: 2935
Accuracy: 97.9%

Total true positives: 2806 (0.9359573048699132)
Correct positives: 2750
Sensitivity: 98%
Total true negatives: 192 (0.06404269513008673)
Correct negatives: 185
Specificity: 96.35%

Flowers:

Overall:
Total classifications: 2998
Correct classifications: 2918
Accuracy: 97.33%

Total true positives: 1592 (0.5310206804536357)
Correct positives: 1545
Sensitivity: 97.05%

Total true negatives: 1406 (0.46897931954636424)
Correct negatives: 1373
Specificity: 97.65%

Fruits:

Overall:
Total classifications: 2998
Correct classifications: 2904
Accuracy: 96.86%

Total true positives: 796 (0.26551034022681785)
Correct positives: 739
Sensitivity: 92.84%

Total true negatives: 2202 (0.7344896597731821)
Correct negatives: 2165
Specificity: 98.32%

Advertisements

Training the Machines II

We wanted to thank all the Notes from Nature volunteers who helped with Training the Machines I, which focused on Prunus – the cherries, plums, almonds, peaches and nectarines.  We don’t yet have the accuracy of our NFN volunteers compared to a gold standard dataset we created, but we are hard at work on that. We do have reconciliations done, which provide some information about how consistent everyone was, and here are some of those results:  For flowers, there were 2775 strict matches (all three agreed) and 223 majority matches.  For fruits, there were 2,663 strict matches and 335 majority rule matches.  And for leaves unfolded, there were 2,589 matches and 409 majority rule matches.  These are mostly encouraging results!

We also are going to be looking at other interesting questions with the results from this expedition, including some trends in accuracy over time — does scoring more samples mean people get better at this? Or maybe fatigue sets in? We also want to look at accuracy over different species — some might be more challenging than others (we are looking at you, desert almond!). We also see if strict or majority are more likely to be right or wrong. Anyway, we have some key hypotheses to test and we are working on those results and will report more. And of course all this work will be feeding into approaches to scale up machine learning, which we think is exciting – and which we also will have more to share with you soon.

We also need to ask for your help once more, this time on the plant group Acer, the maples. Acer, like Prunus, is well studied for phenology, and has an impressive historical and current record of observation. But Acer can also be challenging (talking to _you_, box elders!). So please pay close attention the help guides which can really help you out here.

Also, your work is really helping out @naturalista, who will be working on a dissertation chapter and papers comparing these results, so thank you so much for the help, and hopefully you will really enjoy this expedition focused on maples, an iconic shade tree that is especially valued in the heat of the summer.

— Rob Guralnick, University of Florida

Happy Anniversary!

Happy Anniversary to Notes from Nature!

Today — June 16th, 2019 — Notes from Nature is celebrating 6 years!

Why June 16th? For a little history, NfN 1.0 launched on Earth Day (April 22nd) 2013. We upgraded to NfN 2.0 on June 16th, 2016 and NfN 3.0 on May 9th, 2019. Although our history is a bit complicated since we have made major platform changes every 3 years or so, we have made it a tradition to celebrate our anniversary on June 16th.

Please help us celebrate 6 years of Notes from Nature by doing a few transcriptions today! June is a slow time of year and we sure could use some help finishing off those remaining NfN 2.0 expeditions.

With gratitude,

The Notes from Nature Team

p.s. Happy Father’s Day to all of the fathers out there! Thanks for your support!

A challenge!

We are very excited to launch a new Project related to the plants of Arkansas, U.S.A. However, there is some unfinished business on NfN 2.0. The expedition is called Plants of Arkansas: Digitizing Hendrix College’s Collection. It’s currently at 26% complete and we’d love to get it finished ASAP.

Please take a few minutes and do 5 (or more!) classifications to help us finish this one off.

You can find out more about the exciting things going on in Arkansas at the SUPERB and ACBC Facebook pages.

— The Notes from Nature Team

 

Notes from Nature 3.0!

The time has come. We encourage you all to come on over to Notes from Nature and check out our recent upgrade. We have even started a specific Talk board for you to post your comments and questions.

Below are a few important aspects of the upgrade.

The Notes from Nature page will be organized around Projects. This will give providers new ways to add content and engage with volunteers. Today we launch with two new Projects from our old friends at WeDigFLPlants and Calbug! These are similar to the groups that we used before, but give data providers many more options for serving their content.

All the remaining Notes from Nature expeditions from 2.0 can be accessed through the “Notes From Nature” project link until they are all complete.

The custom Notes from Nature Field Book should function as expected and should contain all your classifications from NfN 2.0.

You may notice that the new page does not contain an Organization wide statistics page or a completed expeditions page. Both of these will be added back in the coming months. In the meantime, you can see some limited Organization-wide and full Project-wide statistics on the respective landing pages and through the links found there.

You will also note that each Project has its own Talk page. While we may utilize this option more in the future, for now these will only contain Notes posts related to subjects within that Project. The main NfN board will remain our primary discussion place.

Since there is room for more content on the Project pages, the expedition cards that we used in 2.0 have been replaced with a simple button that only contains the name of the expedition. This will be the case for the time being, but we do plan to revamp these buttons again in the coming months.

The Organization and Projects links have moved around a bit, so be sure to look at the menu bar, but also the lower right link area of each landing page.

— The Notes from Nature Team

Coming soon (NfN 3.0)

Well, the time has just about arrived. If you are active on the Zooniverse in general or NfN Talk you have heard us mention an upcoming update to NfN that has been in the works for some time. We have been functioning on the NfN 2.0 platform (aka Panoptes) for almost three years now. We launched NfN 2.0 on June 16th, 2016. With over 1 million classifications and close to 8,000 registered volunteers we have outgrown that platform and are just about ready to move to NfN 3.0! There are lots of details to work out, but we plan to launch this week and hopefully as soon as this Wednesday.

There are some projects such as Snapshot Safari that have are utilizing this new platform that is called Organizations. You’ll notice that the look and feel is largely the same, but the site will now feature Projects that allow for more specific content to be added by our data providers.

It is worth noting that there will be no disruption of service for NfN during the upgrade and that all current expeditions on NfN 2.0 will be accessible and will be completed before, during and after the transition if needed.

Thanks to all our data partner providers and volunteers for making NfN such a success. We also want to thank the Zooniverse team for all their support and development!

— The NfN Team

Happy Earth Day 2019!

Earth_Day_Flag.png

Dear NfN Community,

Happy Earth Day 2019!

Today we want to honor our wonderful NfN community for partnering with us to conserve and make available knowledge about the natural world. The NfN project gives you the opportunity to make a scientifically important contribution towards that goal every single day. We are so very thankful for your contributions today and every day!

Every transcription that is completed brings us closer to filling gaps in our knowledge of global biodiversity and our natural heritage.

— The Notes from Nature Team

 

%d bloggers like this: