Tag Archives: Learning

Did you know that bumblebees find efficient ways to move a ball to a target?

Have you ever watched a bumblebee collecting nectar from flowers? At first sight, this does not seem like a very difficult task, requiring sophisticated cognitive skills.

However, bumblebees have an amazing ability to learn. Not only where to find the nectar, but also how to perform human-designed tricks that require behaviours not needed in nature. For example, bumblebees have learned to pull strings, pull caps aside, and rotate disks to get to a reward.

Not so long ago, scientists (and bumblebees) have taken it a step further. The scientists, in various ways, demonstrated to their subjects that moving a ball into a target (a circle drawn on the experimental platform) led to a reward. The bumblebees quickly learned to do so themselves, and even improved their tactic to be more efficient than those demonstrated.

Teachers

First, the researchers trained the bumblebees to move a wooden ball, larger than them, to a target – the demonstrator was an artificial insect (on a stick held by the experimenter) pushing the ball. The bumblebees quickly got the idea, but preferred pulling the ball while moving backwards over pushing it. These bumblebees were the demonstrators in the next phase.

Students

In the main experiment, new, untrained bumblebees were divided into three groups. Each insect in the first group could watch another bumblebee dragging one of three available balls to a target and they both received a reward – a drop of sugar water. Insects in the second group saw a ball that moved “by itself” to the target (with help of a researcher-directed magnet under the platform). When the ball reached its destination, the watcher would get a reward. Insects in the third group simply found the ball already at the target with the reward next to it. Each demonstration/trial was conducted only three times.

Then the bumblebees were tested without further demonstration and only got a reward if they brought the ball to the target themselves. Virtually all bumblebees that had had a live demonstrator successfully dragged the ball to the target and did so faster than the other two groups. Those that had seen the ball moving by itself completed about 8 out of 10 trials. Finally, those bumblebees that had simply found the ball next to the reward were successful on average in only 3 or 4 of 10 trials and took the longest.

The student has become the master

Interestingly, the bumblebees did not simply copy what they observed early on. The bumblebee-demonstrator and magnet-using scientist always moved the farthest ball to the target. The student bumblebees usually moved the one closest to the target, even if it had a different colour than the one in the demonstration. And it was not a result of closest ball being pushed accidentally to the target, because the bumblebees usually dragged the ball actually being in-between the ball and the target.

To sum up: firstly, the bumblebees quickly learned a new task requiring the use of a tool (ball) and behaviour that has little to do with normal bumblebee foraging. And secondly, they did not blindly copy a previously observed behaviour, but used a more efficient method – moving the ball closest to the target.

Such unusual experiments show how great a capacity for learning and flexibility in problem solving these common insects have.

Rats in action 2 / Szczury w akcji 2

After a handful of attempts (for video of the first trials see here) my rats figured out how to open my LEGO puzzle boxes. It took them the longest to realise how to open the door of the DUPLO house. Tokyo was more interested than Stripe in the whole affair and is better at the ask.


Po paru próbach (film z pierwszych prób możesz naleźć tutaj) moje szczury nauczyły się otwierać pudełka-łamigłówki. Najtrudniejsze okazało się otwieranie drzwi do domku z DUPLO. Tokjo okazała większe zainteresowanie pudełkami i nauczyła się je otwierać szybciej.

Did you know that rats can learn to play hide-and-seek?

Many pets like to play, not only when they are young. But have you ever played hide-and-seek with one? And I’m not talking about only you hiding so that your dog can sniff you out. I’m talking about the game we probably all know from childhood where one person stays in one place, while the other hides and waits to be found by the first one. And afterwards, the roles are reversed!

Unfortunately, it wasn’t me playing hide-and-seek with rats, but researchers from Germany. They trained the rats to play both seeker and hider.

When seeking, rats were locked in a central box and released remotely once the researcher had hidden themselves behind one of several cardboard covers. As soon as the rat found the researcher, they got a reward. Not a tasty treat, but an interaction, like tickling, rough-and-tumble-like play and petting.

When hiding, the rat was placed in an open central box. The researcher sat next to it, loudly counting. The rat had 90 second to hide. If it was hidden behind a cover or in one of the boxes and stayed there till the researcher found it, it got a reward of interaction.

The rats were successful at both hiding and seeking and they could easily switch between the roles.

Moreover, they used strategies I recognise from my own childhood, even though they were not explicitly trained to use them!

When seeking rats moved more, checking different locations in a systematic manner and they used visual cues. They also checked previous hiding locations of humans more often (sometimes researchers kept on hiding in the same place) while rarely entering small boxes.

When hiding, their movement was more directed and they changed hiding locations between trials and stayed quiet (also in the ultrasound range they normally use for communication) while hiding.

And it seems that the rats were not playing just to get a reward, but that they actually had fun!

They performed joy jumps, and when found they often cut short the reward interaction, ran away and re-hid, thus prolonging the game and delaying the reward!

I wonder, shall I also train my rats to play hide-and-seek? But I have a feeling that they would rather choose pumpkin seeds as a reward than tickling…

***

You can watch the recording of the game here.

***

Polską wersję tego wpisu możesz znaleźć tutaj.

Did you know that great tits are conformists and can form traditions?

Probably everyone in Europe is familiar with this little bird – great tit. And the winter is a great period to observe them as they often come to birdfeeders in gardens and on balconies. They usually come in pairs or small groups.

It is known that these birds are innovative but they can also learn from each other and they seem to follow the behaviour of majority of a population (group of birds that live in one area). Therefore, showing conformity.

At least when it comes to opening puzzle boxes. A few years ago, researchers in the UK trained a couple of wild great tits from different locations to open a puzzle box by either sliding its door to the left or to the right (one bird learned only one method)*.

Then they released them into the wild and provided puzzle boxes filled with mealworms (a delicacy for the great tits) that could open either way (both to the left and to the right).

Within 20 day most of the birds knew how to open the boxes and it seems that they learned it from the trained individuals, as in populations without trained birds much fewer individuals managed to get to the worms.

But what is more interesting, most of the birds in a given area used the method that was used by the trained bird (left or right slide) even if the other way was equally difficult and rewarding. This shows social learning from others.

But here comes even a more amazing finding. There were birds that actually used both ways of opening the box (some probably learned the less common variant individually, by trial and error). However, most of these individuals still preferred the most common behaviour. And some of them even switched from the less common to more common behaviour. But never the other way round.

If immigrant birds came from areas with different traditions, most of them changed their behaviour to match the locals (A reader who is themselves an immigrant will probably understand this).

When researchers returned the next year, they saw that the local traditions were even more pronounced, with fewer birds using uncommon method to open the box.

All these observations show that not only humans follow societal norms and have traditions (group-specific, socially learned, and often arbitrary behaviour) that they learn through conformism.

Although, like in human populations, there were still few birds that did not follow societal norms and just slid that door against local tradition.

* How do you train the bird to open the box the way you want? Well, block the door so it can open in only one way. The training part is quite simple (in case you want to train your own pets – please do not capture wild birds!). First show the birds an open box with worms inside that they can just pick up. Then, each time close the door more and more until the birds can reliably open it even if it’s fully closed (great tits learned that in 4 days).


Photo of great tit by Petr Ganaj from Pexels.com


Polską wersję tego wpisu możesz znaleźć tutaj.