Subject Areas

CPAL 2027 welcomes contributions in which a clear parsimony principle is central. The following list is illustrative rather than exhaustive.

Theory and Foundations

Sparsity, structured sparsity, low rank, subspaces, manifolds, tensors, graphs, latent-variable models, and other low-dimensional structures.
Statistical and computational foundations of representation and feature learning.
Identifiability, generalization, robustness, implicit bias, and statistical–computational tradeoffs.
Symmetry, invariance, equivariance, causality, modularity, and compositional structure.
Information-theoretic, minimum-description-length, and compression-based views of learning.
Theoretical neuroscience, cognitive science, and biologically inspired mechanisms for parsimonious learning.

Methods and Models

Sparse coding, dictionary learning, matrix and tensor methods, inverse problems, and structured recovery.
Optimization and feature-learning dynamics that discover or induce parsimonious structure.
Interpretable and structured neural architectures, including algorithm unrolling.
Pruning, quantization, distillation, low-rank adaptation, modular networks, mixtures of experts, and conditional computation.
Data-, parameter-, memory-, energy-, and compute-efficient training and inference.
Parsimonious foundation, generative, multimodal, federated, continual, and agentic learning.

Systems, Data, and Applications

Hardware–algorithm and software–system co-design for structured or sparse computation.
Benchmarks, datasets, and metrics that reveal or evaluate parsimony and accuracy–efficiency tradeoffs.
Parsimonious learning for perception, action, reasoning, robotics, and control.
Scientific machine learning, signal and image processing, neuroscience, biology, medicine, engineering, and social science.
Resource-constrained, distributed, networked, and real-world intelligent systems.

Submissions proposing a new notion of parsimony are especially welcome when the connection is clearly motivated. Authors who are uncertain about fit may contact the Program Chairs at pcs@cpal.cc.