I Saw This One Paper Where They Made An Artificial Neural Network Based On The Actual Neural Architecture

I saw this one paper where they made an artificial neural network based on the actual neural architecture of the fruit fly and trained it on pictures of flies to show that 1. individual fruit flies are visually distinct 2. they are probably able to differentiate between each other visually despite their vision being terrible. And as a comparison they had a bunch of experienced fly scientists (aka “flyentists”) try to identify the same pictures of flies and they failed miserably which I thought was really funny

This ability to re-identify flies across days opens experimental possibilities, especially considering that this performance was achieved with static images (16fps yields around a thousand estimates of ID per minute, allowing high confidence in the parsimonious correct identification). This is in contrast to the human ability to re-identify flies, which at low resolutions is barely better than chance.

Clearly, all models can learn to re-identify flies to some extent, underscoring the individual-level variation in D. melanogaster. Re-identifying flies is in fact easier for DCNs than CIFAR10 (at least with centred images of flies acquired at the same distance). Even the model that rivals, in some sense, the representational performance of humans does ten times better than humans. Why humans can’t tell one fly from another is not clear. Regardless of whether it was evolutionarily beneficial to discriminate individual flies, humans do have incredible pattern detection abilities. It may simply be a lack of experience (although we attempted to address this by only using experienced Drosophila researchers as volunteers) or a more cryptic pattern-recognition ‘blind-spot’ of humans. In either case, these findings should spur new experiments to further understand the mechanisms of human vision and experience and how they fail in this case.

these CRINGE scientists FAILED to identify flies that all our models could smho 🙄😤

More Posts from Phoronopsis and Others

2 months ago

“Measuring sea cucumber body dimensions and weight and determining their relationship is notoriously difficult.” — Prescott, Zhou & Prasetyo 2015

“Tagging sea cucumbers is notoriously difficult because of their plastic nature and autolysis capacities.” — Gianasi, Verkaik, Hamel & Mercier 2015

“Nevertheless, marking and tracking sea cucumbers is notoriously difficult and represents a serious challenge.” — Rodríguez-Barreras, Lopéz-Morell & Sabat 2016

“Obtaining accurate but non-destructive mass and morphology measurements of holothuroids is notoriously difficult because they readily change shape and retain water in their body cavity.” — Munger, Watkins, Dunic & Côté 2023

the notoriously difficult cucumber

A photo of the elephant trunkfish, Holothuria fuscopunctata, pictured underwater. It is a golden-brown sea cucumber with small dark spots and a light underside, with a shape resembling that of a baguette.

image by Amaury Durbano


Tags
6 months ago

This crab is under construction! Read more on the Aquarium's website. 🦀🦺

Caring for Crustaceans with Creativity
aquariumofpacific.org
An aquarist cares for crabs in need of new shells in a new, inventive way, by using non-toxic epoxy, molding, and paint.

Tags
1 week ago

Conch snails actually do have some of their own tricks up their shells— their foot bears a sharpened operculum that they use to push themselves around much faster than a lot of slow predators (including cone snails) can move, or even to fight back. It's believed that their high-resolution vision, which is some of the best among all known gastropods, allows them to detect and react to predators in advance (source 1, 2)

Here's a video of a conch snail in action:

How Are Conchs Even Real

How are conchs even real


Tags
6 months ago
Thinking About That Time Scientists Put Tracking Collars On Jellyfish

thinking about that time scientists put tracking collars on jellyfish


Tags
7 months ago

During the storm, I've had a wasp sheltering on my window.

It's been two days now, and she's still there, so I gave her some honey

During The Storm, I've Had A Wasp Sheltering On My Window.
During The Storm, I've Had A Wasp Sheltering On My Window.

Lookit her! slurping away!

@onenicebugperday


Tags
1 month ago

If your girl has

bulbous eyes

piercing-sucking mouthparts (beak)

raptorial legs

cogwheel-like structure

that’s not your girl that’s wheel bug!

A photo (credit: Joe Boggs) of an adult wheel bug standing on a plant. It is a gray, sturdy-looking insect with all the features mentioned above highlighted with labeled arrows.

(photo from this article)


Tags
3 months ago

we all know people who go out of their way to be rude on bug appreciation posts are annoying as heck but sometimes they manage to read the room so absurdly poorly that it's just funny. You'll see a photo with 200 notes by someone called "flylover4ever" with the caption "look at this beautiful blowfly I found on my morning bug hunt 😊" and every comment note and tag is something like "look at that coloring!" "what beautiful eyes you have 😍" "KISSING HER ON THE TERGAL PLATE" and then there's just one rando person being like "EWWW kill it with fire 🤮". And it's like how did you even get here. are you lost, where did you even come from


Tags
4 months ago

this isn't related to cephalopods at all, but re: bilateral symmetry, there's actually some fairly recent (like, last 15 years) evidence that cnidarians actually evolved from bilaterally symmetric ancestors! Contrary to popular belief, a lot of sea anemones and coral polyps, though externally radial, actually have a bilateral digestive system. This could be a case of convergent evolution, but what's really remarkable is that embryonically, cnidarians develop this bilateral symmetry the same way as bilaterians, *using the same regulatory genetic pathways*! The polyp body plan is considered to be the ancestral state for cnidarians, while the "simplified" swimming jellyfish body plan probably is a secondary development, as suggested by the cnidarian fossil record and evidence of loss of body patterning genes in jellyfish. Together, this suggests the really exciting hypothesis that bilateral symmetry is actually the original ancestral state for the common ancestor of cnidarians and bilaterians, and rather than bilaterians being the ones who made the innovation of bilateral symmetry from radial, cnidarians would be the group that altered their symmetry instead! Sources: 1, 2, 3

(Echinoderms like starfish also are secondarily radial; they have bilateral larvae who undergo a very weird metamorphosis into their pentaradial adult form. Search "brachiolaria", "pluteus larva", or "auricularia larva" for pics!)

It's also a common misconception that cnidarians don't have central nervous systems. They're often said to have merely diffuse "nerve nets," and they do, but they also have a condensed ring-shaped nerve that integrates signals from across the body, basically constituting a central nervous system in all respects besides not actually having a singular "brain". In particular, there's been a lot of research into the nervous systems of box jellyfish, which are probably the most specialized among cnidarians due to their unique possession of true image-forming eyes, which they use for navigating both long and short distances as well detecting prey. One study from just last year even found evidence that box jellies display associative learning! Sources: 1, 2, 3, 4

(I believe there's also been research into the learning capabilities of echinoderms but I'm not as familiar with the literature. I know starfish do actually have image-forming eyes on their arms, which at least one species uses to navigate, though they also definitely do a lot of smelling and stuff as well. Source: 1)

Imagine yourself submerged in the prehistoric ocean. There are no fish, instead the only life forms consist of feather-like sessile organisms that sit on the seabed, filtering the current. The early organisms that evolved out of this, such as Jellyfish and Starfish, had radial anatomy. Their body structure entails a central axis from which you can split everything else. These bodies are simple, not designed for active mobility, lacking a ‘forwards’ or ‘backwards’. They didn’t even have eyes, instead interacting with and responding to the world via photoreceptive cells. What emerged from this were two developments: the evolution of complex eyes and the emergence of bilateral anatomy in early vertebrates and arthropods. In contrast to radial anatomy, bilateral anatomy entails an organism that can be split down the middle with rough symmetry. This is to say that they are built for direction. A body that is built for mobility entails significantly more complex behaviour behind its operation. Behaviour, in this sense, also becomes significantly more directed. These creatures now living in the ocean or on the sea-floor now begin to directly interact with one another. The mechanisms facilitating this interaction become pretty apparent in the fossil record; eyes, claws and antennae. The evolutionary consequences of this are the emergence of a complex nervous system alongside the presence of predation and, as Godfrey-Smith puts it “[From this point on] The mind evolved in response to other minds”. 


Tags
2 months ago

The first image linked is not actually a priapulid but a sea cucumber in its spawning posture! It was misidentified on iNaturalist and went viral before it was corrected— see the original observation here. (It gets kinda heated which I think is kinda funny. Penis worms are serious business!) I have always said before that I want internet fame specifically for two reasons: to make PSAs about Anomalocaris’s head carapace which everybody always leaves out of drawings because of that one inaccurate museum model, and about the incorrectly identified sea cucumber photo about which is now like the first image result you get when searching for penis worms and is my NEMESIS 😠 (the misinformation, not the photo or the sea cucumber, those are great)

For all the worm fans— priapulids are super easy to identify; there are as of the time of writing only 22 recognized species, and for many of them the only photos of them are from articles in scientific journals. Over half the species are microscopic, and the macroscopic ones are mainly found in polar regions, often in the deep sea, where they are usually burrowed in sediment and thus are little-encountered by people. The only one of them that is commonly photographed (and studied) is Priapulus caudatus, which is broadly found across the northern northern hemisphere even in shallow waters and I think probably has to be the most accessible species in general. They look like this:

A live specimen of the priapulid worm Priapulus caudatus, pictured in a petri dish. It is a pinkish-tan worm with a bulbous, spined proboscis and a branching, feathery tail.

image by Thomas Trott

This species accounts for probably 99% of the images of priapulids out there, and its relatives look rather similar, such as its southern hemisphere counterpart Priapulus tuberculatospinosus or the two-tailed species Priapulopsis bicaudatus. The intricate, feathery tails (referred to in the literature as “caudal appendages”) are probably the most distinctive feature of this group; they are believed to be involved in respiration, though as with many things about the phylum it is not known for certain. (See this recent paper for a review of macroscopic priapulid morphology.) In the zoomed-out photos of that sea cucumber you can see on the iNat page, it lacks a tail which is a dead giveaway that it is not any of these; also note that while it has some longitudinal striations along what sorta looks like a proboscis, they don’t actually bear any teeth! The spined, toothed proboscides of priapulids are indeed super cool and are their most distinctive feature setting them apart from other proboscis-bearing worms like peanut worms or spoon worms, which are often also misidentified online as priapulids. A fun fact is that the shape of their teeth varies across species in a way that appears to be closely correlated with their diet, see this paper for a neat study that uses tooth shapes to examine the different ecological niches occupied by extant priapulids and their Cambrian relatives!

The only other macroscopic priapulids that don’t look much like Priapulus are the two species Halicryptus spinulosus and Halicryptus higginsi, the latter of which I believe there are literally like two full-body photos in existence of it, one of which is from a login-walled journal article from 1999 and the other of which is one of the specimens from that 1999 article photographed after 25 years preserved in a museum. There’s a decent number of photos floating around of H. spinulosus (though still not as many as P. caudatus); they look like this:

A group of specimens of the priapulid worm Halicryptus spinulosus, photographed in a petri dish. They are small, cylindrical yellowish worms with visible spines on their short proboscis.

image by Claude Nozères

As you can see, Halicryptus lack tails and have a much less prominent proboscis than Priapulus and its relatives, which you can only see the spines of on the very tip; H. spinulosus in particular has a rather short and small body that distinguishes it a lot, while H. higginsi is the largest known species of priapulid in the world (see this paper for a review of both of them). They’re maybe less distinctive-looking but idk, I don’t know off the top of my head if there’s super anything else you would mistake them for, and images of them are pretty uncommon anyway. In any case as far as macroscopic priapulids go, these are the only ones you have to look out for; if you’ve got those down you’re all set! As stated before, most priapulid species are actually microscopic; just for fun here’s the tropical meiobenthic species Tubiluchus corallicola:

A preserved specimen of the priapulid worm Tubiluchus corallicola. It is a microscopic worm with translucent skin revealing yellowish internal organs and a long, slender tail.

image by Museum of Comparative Zoology, Harvard University

look at that squiggly tail!

And yeah in conclusion priapulids are super cool and underrated and I wish there were more people paying attention to them; there’s soooo many neglected taxa that we’re still only just discovering basic aspects of their biology and priapulids are one of them! If you want to see their amazing extensible proboscis in action, linked below is by far the best priapulid video out there, I highly recommend it. And most of all remember everybody THAT PHOTO IS A FRICKING SEA CUCUMBER, NOT EVERY WORM THAT LOOKS LIKE A PENIS IS A PENIS WORM AAAAAAA 😭😭😭

Can't believe any real animal has teeth as awesome as penis worms have.

Can't Believe Any Real Animal Has Teeth As Awesome As Penis Worms Have.
Can't Believe Any Real Animal Has Teeth As Awesome As Penis Worms Have.

They are meat eaters :)


Tags
6 months ago

bugs are always cleaning their damn antennas


Tags
Loading...
End of content
No more pages to load
  • jacqueslasseau
    jacqueslasseau liked this · 3 months ago
  • peelerofpotatoes
    peelerofpotatoes liked this · 7 months ago
  • glurblr
    glurblr liked this · 7 months ago
  • petersillie
    petersillie liked this · 7 months ago
  • phoronopsis
    phoronopsis reblogged this · 7 months ago
phoronopsis - actinotroch
actinotroch

they/she ✩ I like space and invertebrates

45 posts

Explore Tumblr Blog
Search Through Tumblr Tags