Rate-distortion optimization for transformer inference
Split computing for language models, extending the theory of usable information.
Lossy common information in a learnable Gray-Wyner Network
Isolate the common information between two dependent computer vision tasks.
Rate-distortion theory in coding for machines and its applications
Theoretical considerations and evaluation of split and distillation points.
Towards task-compatible compressible representations
Task reconstruction loss acts as a regularizer, increasing rate-distortion performance in coding for humans and machines.
Base layer efficiency in scalable human-machine coding
Improving the shared channel in coding for machines (CfM).
Conditional and residual methods in scalable coding for humans and machines
A comparison between conditional and residual entropy codecs for a two-channel systems of tasks with nested information.
An architecture for accelerated large-scale inference of transformer-based language models
Unified batch and online transformer inference.
Graph representation learning network via adaptive sampling
Graph representations using a learnable attention mechanism to sample the neighbourhood of a graph.
DENS: A dataset for multi-class emotion analysis
Dataset for emotion classification of long-form narratives.
Exploring multilingual syntactic sentence representations
Sentence embeddings augmented by universal parts-of-speech tags evaluated on low-resource languages.
Unsupervised aspect extraction from free-form conversations
Dictionary-based approach for the extraction of “aspect-of” relationships.
Best practices for convolutional neural networks applied to object recognition in images
Evaluate the performance impact of optimization algorithms, activation functions, dropout, and maxout networks, in CNNs.
A comparison of neural network training methods for text classification
Benchmark of stochastic gradient descent and Nesterov’s accelerated gradient for text classification.