Bruno Gavranović
@bgavran.bsky.social
Categorical Deep Learning.
I do get a burst of inspiration every now and then. But these are few and far in-between.
November 3, 2025 at 7:23 PM
I do get a burst of inspiration every now and then. But these are few and far in-between.
Quite hoping it gets back - those have been good days, and I often think about how I wouldn't have become the mathematician and computer scientist I am today if it were not for those days
November 3, 2025 at 7:23 PM
Quite hoping it gets back - those have been good days, and I often think about how I wouldn't have become the mathematician and computer scientist I am today if it were not for those days
Is there a container interpretation in the case of an arbitrary category S?
November 2, 2025 at 10:41 PM
Is there a container interpretation in the case of an arbitrary category S?
This looks like it's a useful resource for thinking about this, though on the first glance it doesn't include anything about tree-shaped arrays, which I'd be very interested in
September 22, 2025 at 8:49 PM
This looks like it's a useful resource for thinking about this, though on the first glance it doesn't include anything about tree-shaped arrays, which I'd be very interested in
I'd been wondering myself about the foundations of the question of 'coordinate transformation' of the logical representation of multi-dimensional arrays or of general inductive types into a physical, 1d linear representation.
September 22, 2025 at 8:49 PM
I'd been wondering myself about the foundations of the question of 'coordinate transformation' of the logical representation of multi-dimensional arrays or of general inductive types into a physical, 1d linear representation.
TLDR; insertion sort is dual to selection sort
September 18, 2025 at 2:23 PM
TLDR; insertion sort is dual to selection sort
"...duality among sorting algorithms: insertion sorts are dual to selection sorts."
First time I hear about this!
First time I hear about this!
September 18, 2025 at 2:23 PM
"...duality among sorting algorithms: insertion sorts are dual to selection sorts."
First time I hear about this!
First time I hear about this!
And this is the part where I conjecture that it has to be a certain kind of a fixpoint. But when it comes to deducing whether we're looking at
\mu(Ext(C)) or \nu(Ext(C)) I am actually not sure...
\mu(Ext(C)) or \nu(Ext(C)) I am actually not sure...
September 16, 2025 at 1:17 PM
And this is the part where I conjecture that it has to be a certain kind of a fixpoint. But when it comes to deducing whether we're looking at
\mu(Ext(C)) or \nu(Ext(C)) I am actually not sure...
\mu(Ext(C)) or \nu(Ext(C)) I am actually not sure...
The natural way to formalise such an infinite product is by trying to compute the final coalgebra of the (- o C) functor, and see whether that gets us anything interesting.
September 16, 2025 at 1:16 PM
The natural way to formalise such an infinite product is by trying to compute the final coalgebra of the (- o C) functor, and see whether that gets us anything interesting.
I *think* this ought to hold also when \mu on the RHS is replaced with \nu
September 16, 2025 at 12:59 PM
I *think* this ought to hold also when \mu on the RHS is replaced with \nu
Actually, I don't think anymore we have the kleene star at hand.
I think now I have a working hypothesis which is the following:
\nu(- o C) \cong Const (\mu(Ext(C)))
I think now I have a working hypothesis which is the following:
\nu(- o C) \cong Const (\mu(Ext(C)))
September 16, 2025 at 12:58 PM
Actually, I don't think anymore we have the kleene star at hand.
I think now I have a working hypothesis which is the following:
\nu(- o C) \cong Const (\mu(Ext(C)))
I think now I have a working hypothesis which is the following:
\nu(- o C) \cong Const (\mu(Ext(C)))
With ordinary containers that gets muddled, so I'm looking to add an extra level of refinement. And singly indexed containers seem like the right thing to write some examples of
September 14, 2025 at 4:04 PM
With ordinary containers that gets muddled, so I'm looking to add an extra level of refinement. And singly indexed containers seem like the right thing to write some examples of
With the current world events unfolding, science might become more and more of a luxury we can't afford until the situation is normalised
September 14, 2025 at 12:45 PM
With the current world events unfolding, science might become more and more of a luxury we can't afford until the situation is normalised
I would highly doubt that indexed containers would be in that book, given that I'm here talking about searching through *research articles* to find examples of a particular concepts
September 8, 2025 at 9:25 PM
I would highly doubt that indexed containers would be in that book, given that I'm here talking about searching through *research articles* to find examples of a particular concepts
Interestingly, a few days ago I've been wondering whether I should purchase this book.
I ended up wondering how much it has kept up with the times - it sounds like until 2012 that was the case, but do you know if that changed after?
I ended up wondering how much it has kept up with the times - it sounds like until 2012 that was the case, but do you know if that changed after?
September 8, 2025 at 9:25 PM
Interestingly, a few days ago I've been wondering whether I should purchase this book.
I ended up wondering how much it has kept up with the times - it sounds like until 2012 that was the case, but do you know if that changed after?
I ended up wondering how much it has kept up with the times - it sounds like until 2012 that was the case, but do you know if that changed after?
Ideally I'd find something like the list of examples I made for containers here
( https://github.com/bgavran/TensorType/blob/main/src/Data/Container/Object/Instances.idr#L10 ) but seeing even an example or two concretely would be useful.
( https://github.com/bgavran/TensorType/blob/main/src/Data/Container/Object/Instances.idr#L10 ) but seeing even an example or two concretely would be useful.
TensorType/src/Data/Container/Object/Instances.idr at main · bgavran/TensorType
Framework for type-safe pure functional tensor processing, supporting non-cubical tensors - bgavran/TensorType
github.com
September 8, 2025 at 6:46 PM
Ideally I'd find something like the list of examples I made for containers here
( https://github.com/bgavran/TensorType/blob/main/src/Data/Container/Object/Instances.idr#L10 ) but seeing even an example or two concretely would be useful.
( https://github.com/bgavran/TensorType/blob/main/src/Data/Container/Object/Instances.idr#L10 ) but seeing even an example or two concretely would be useful.
But I realised that, unlike with ordinary containers - for which there's numerous examples implemented in say Agda/Idris, I'm unable to find concrete examples of indexed containers in code.
September 8, 2025 at 6:46 PM
But I realised that, unlike with ordinary containers - for which there's numerous examples implemented in say Agda/Idris, I'm unable to find concrete examples of indexed containers in code.