11 citations · 24 across the 13 of their papers we have counts for
13 papers
You may implement this later: Cofunctors as partial implementations
Vincent Wang-Maścianica
A functor is a familiar model of an implementation, where every operation in a specification is assigned a concrete instantiation at the outset. But some tasks are less eager: we o…
Graphical einops: bridging tensor networks and computation graphs
Vincent Wang-Maścianica, Nikhil Khatri
Architecture diagrams are ubiquitous in deep learning, but they are usually only representational: the tensor-program identities they suggest are still proved by prose and tensor-a…
Full-Stack Alignment: Co-Aligning AI and Institutions with Thick Models of Value
Joe Edelman, Tan Zhi-Xuan, Ryan Lowe +30
Beneficial societal outcomes cannot be guaranteed by aligning individual AI systems with the intentions of their operators or users. Even an AI system that is perfectly aligned to…
High schoolers excel at Oxford quantum course using pictorial mathematics
Bob Coecke, Aleks Kissinger, Stefano Gogioso +12
We are at the dawn of the second quantum revolution, where our ability to create and control individual quantum systems is poised to drive transformative advancements in basic scie…
Theory-Grounded Evaluation of Human-Like Fallacy Patterns in LLM Reasoning
Andrew Keenan Richardson, Ryan Othniel Kearns, Sean Moss +2
We study logical reasoning in language models by asking whether their errors follow established human fallacy patterns. Using the Erotetic Theory of Reasoning (ETR) and its open-so…
Making the quantum world accessible to young learners through Quantum Picturalism: An experimental study
Selma Dündar-Coecke, Caterina Puca, Lia Yeh +12
The educational value of a fully diagrammatic approach in a scientific field has never been explored. We present Quantum Picturalism (QPic), an entirely diagrammatic formalism for…
On the Anatomy of Attention
Nikhil Khatri, Tuomas Laakkonen, Jonathon Liu +1
We introduce a category-theoretic diagrammatic formalism in order to systematically relate and reason about machine learning models. Our diagrams present architectures intuitively…
A Pattern Language for Machine Learning Tasks
Benjamin Rodatz, Ian Fan, Tuomas Laakkonen +3
We formalise the essential data of objective functions as equality constraints on composites of learners. We call these constraints "tasks", and we investigate the idealised view t…
Constructor Theory as Process Theory
Stefano Gogioso, Vincent Wang-Maścianica, Muhammad Hamza Waseem +2
Constructor theory is a meta-theoretic approach that seeks to characterise concrete theories of physics in terms of the (im)possibility to implement certain abstract "tasks" by mea…
Humans in Humans Out: On GPT Converging Toward Common Sense in both Success and Failure
Philipp Koralus, Vincent Wang-Maścianica
Increase in computational scale and fine-tuning has seen a dramatic improvement in the quality of outputs of large language models (LLMs) like GPT. Given that both GPT-3 and GPT-4…
Distilling Text into Circuits
Vincent Wang-Mascianica, Jonathon Liu, Bob Coecke
This paper concerns the structure of meanings within natural language. Earlier, a framework named DisCoCirc was sketched that (1) is compositional and distributional (a.k.a. vector…
Language-independence of DisCoCirc's Text Circuits: English and Urdu
Muhammad Hamza Waseem, Jonathon Liu, Vincent Wang-Maścianica +1
DisCoCirc is a newly proposed framework for representing the grammar and semantics of texts using compositional, generative circuits. While it constitutes a development of the Cate…
Talking Space: inference from spatial linguistic meanings
Vincent Wang-Mascianica, Bob Coecke
This paper concerns the intersection of natural language and the physical space around us in which we live, that we observe and/or imagine things within. Many important features of…