7 papers
Identifying Informative Environments for Cognition Parameter Inference via Bayesian Experimental Design
Manisha Dubey, Rimvydas Rubavicius, N. Siddharth +1
Computational cognitive modeling seeks to infer latent cognitive mechanisms underlying observed behavior. Bayesian inverse planning provides a principled framework for such inferen…
AIMing for Standardised Explainability Evaluation in GNNs: A Framework and Case Study on Graph Kernel Networks
Magdalena Proszewska, N. Siddharth
Graph Neural Networks (GNNs) have advanced significantly in handling graph-structured data, but a comprehensive framework for evaluating explainability remains lacking. Existing ev…
Towards Human Motion World Models via Executable Behaviour Representations
Rimvydas Rubavicius, Manisha Dubey, N. Siddharth +1
Human motion world models should capture motion's intentionality by being executable: adaptable to different actions and capable of assessing motion quality. To achieve this, we in…
On Designing Diffusion Autoencoders for Efficient Generation and Representation Learning
Magdalena Proszewska, Nikolay Malkin, N. Siddharth
Diffusion autoencoders (DAs) are variants of diffusion generative models that use an input-dependent latent variable to capture representations alongside the diffusion process. The…
Learning High-Frequency Functions Made Easy with Sinusoidal Positional Encoding
Chuanhao Sun, Zhihang Yuan, Kai Xu +4
Fourier features based positional encoding (PE) is commonly used in machine learning tasks that involve learning high-frequency features from low-dimensional inputs, such as 3D vie…
Banyan: Improved Representation Learning with Explicit Structure
Mattia Opper, N. Siddharth
We present Banyan, a model that efficiently learns semantic representations by leveraging explicit hierarchical structure. While transformers excel at scale, they struggle in low-r…
On the effect of curriculum learning with developmental data for grammar acquisition
Mattia Opper, J. Morrison, N. Siddharth
This work explores the degree to which grammar acquisition is driven by language `simplicity' and the source modality (speech vs. text) of data. Using BabyBERTa as a probe, we find…