Greedy Poisson Rejection Sampling: An Efficient Algorithm for One-Shot Channel Simulation
Greedy Poisson rejection sampling (GPRS) is a new rejection sampling algorithm that can efficiently simulate samples from a target distribution Q using a proposal distribution P. GPRS achieves optimal average runtime and codelength for one-dimensional channel simulation problems where the target-proposal density ratio is unimodal.
RECOMBINER: Robust and Enhanced Compression with Bayesian Implicit Neural Representations
RECOMBINER verbessert die Kompression durch lineare Reparametrisierung, Positionscodierungen und hierarchische Modelle.
RECOMBINER: Robust and Enhanced Compression with Bayesian Implicit Neural Representations
RECOMBINER addresses limitations of COMBINER by enhancing variational approximation, adding positional encodings, and utilizing hierarchical models for robust data compression.
Corra: Correlation-Aware Column Compression
Exploiting correlations in data for efficient column compression.
SlimSAM: A Data-Efficient Method for SAM Compression
SlimSAM introduces a novel data-efficient compression method for Segment Anything Model (SAM) that achieves superior performance with minimal training data.
RECOMBINER: Robust and Enhanced Compression with Bayesian Implicit Neural Representations at ICLR 2024
Implicit Neural Representation-based data compression methods are enhanced by RECOMBINER through innovative solutions, achieving competitive results across various data modalities.
SlimSAM: Data-Efficient Compression Method for SAM Models
SlimSAM introduces a data-efficient compression method for SAM models, achieving superior performance with minimal training data.
BAT-LZ Compression Algorithm: Bounded Access Time Lempel-Ziv Variant
Introducing BAT-LZ, a variant of the Lempel-Ziv compression algorithm with bounded access time for efficient text parsing.
Exploring Height-Bounded Lempel-Ziv Encodings for Fast Access
Height-bounded LZ encodings offer efficient access to text positions with minimal space requirements.
RECOMBINER: Robust and Enhanced Compression with Bayesian Implicit Neural Representations
The author presents RECOMBINER as a method to address limitations in COMBINER by enhancing compression through linear reparameterization, positional encodings, and hierarchical models.
