DE-PSGLD is the first decentralized MCMC sampler for constrained convex domains that converges to a regularized Gibbs distribution with explicit 2-Wasserstein bounds for agents and network averages.
hub
[WY20] Hao Wang and Dit-Yan Yeung
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
hub tools
citation-role summary
dataset 1
citation-polarity summary
verdicts
UNVERDICTED 13roles
dataset 1polarities
use dataset 1representative citing papers
Fractionally supervised classification is extended to maxima nominated samples via a new latent representation of the observed maximum and the unseen set composition, producing a valid EM algorithm and weighted-likelihood procedure.
Split-MoPE integrates split learning with predefined-expert routing to maximize usable data in vertical federated learning under sample misalignment, delivering state-of-the-art accuracy in one communication round plus built-in robustness and per-sample contribution scores.
HADES selectively encrypts privacy-sensitive features identified by PCA in federated learning, trains hybrid encrypted and plaintext networks, and fuses them to match vanilla FL accuracy with reduced overhead and better privacy.
A survey of on-device learning in TinyML organized by distribution change regimes, highlighting influences on applications, hardware, and solutions plus a gap between benchmarks and deployments.
TDS uses per-tree prediction trajectories to derive instance difficulty scores that rank errors better than prior hardness measures and improve active learning, selective prediction, and Mondrian conformal prediction on tabular data.
MANOJAVAM unifies matrix multiplication and SVD for PCA on FPGA with block-streaming systolic arrays and pipelined Jacobi-CORDIC, delivering up to 22.75x SVD speedup and 42.14x lower energy than an NVIDIA A6000 GPU.
HiSS sampling uses logistic bridging in a Metropolis-within-Gibbs framework to enable transitions between disconnected modes while preserving the target discrete distribution.
XL-MIMO systems with analog combining perform OTA classification via ELM framework achieving over 90% accuracy with few ms latency under rich fading.
The thesis presents Pino, an end-to-end pipeline that supervises reinforcement learning agents with argumentation-based normative advisors, introduces an algorithm for automatic argument extraction, and defines a mitigation strategy for norm avoidance.
A smoothing stochastic gradient descent algorithm is introduced for non-smooth stochastic compositional optimization, achieving 1/T^{1/4} rate for convex cases and similar guarantees under other convexity settings.
Gini MDS replaces Euclidean distance in multidimensional scaling with a rank-and-value-based pseudo-distance controlled by a hyperparameter, claimed to yield more robust embeddings on noisy data than standard MDS.
A unified large deviations analysis is proposed to study acceleration mechanisms in variants of overdamped Langevin Monte Carlo methods, supported by numerical experiments.
citing papers explorer
-
Decentralized Proximal Stochastic Gradient Langevin Dynamics
DE-PSGLD is the first decentralized MCMC sampler for constrained convex domains that converges to a regularized Gibbs distribution with explicit 2-Wasserstein bounds for agents and network averages.
-
Fractionally Supervised Classification with Maxima Nominated Samples
Fractionally supervised classification is extended to maxima nominated samples via a new latent representation of the observed maximum and the unseen set composition, producing a valid EM algorithm and weighted-likelihood procedure.
-
Mixture of Predefined Experts: Maximizing Data Usage on Vertical Federated Learning
Split-MoPE integrates split learning with predefined-expert routing to maximize usable data in vertical federated learning under sample misalignment, delivering state-of-the-art accuracy in one communication round plus built-in robustness and per-sample contribution scores.
-
HADES: Privacy-Preserving Federated Learning via Selective Feature Encryption and Hybrid Model Fusion
HADES selectively encrypts privacy-sensitive features identified by PCA in federated learning, trains hybrid encrypted and plaintext networks, and fuses them to match vanilla FL accuracy with reduced overhead and better privacy.
-
What changes after deployment? A survey on On-device Learning in TinyML
A survey of on-device learning in TinyML organized by distribution change regimes, highlighting influences on applications, hardware, and solutions plus a gap between benchmarks and deployments.
-
Trajectory-Based Difficulty Scoring for Reliable Learning on Tabular Data
TDS uses per-tree prediction trajectories to derive instance difficulty scores that rank errors better than prior hardness measures and improve active learning, selective prediction, and Mondrian conformal prediction on tabular data.
-
MANOJAVAM: A Scalable, Unified FPGA Accelerator for Matrix Multiplication and Singular Value Decomposition in Principal Component Analysis
MANOJAVAM unifies matrix multiplication and SVD for PCA on FPGA with block-streaming systolic arrays and pipelined Jacobi-CORDIC, delivering up to 22.75x SVD speedup and 42.14x lower energy than an NVIDIA A6000 GPU.
-
Slithering Through Gaps: Capturing Discrete Isolated Modes via Logistic Bridging
HiSS sampling uses logistic bridging in a Metropolis-within-Gibbs framework to enable transitions between disconnected modes while preserving the target discrete distribution.
-
Universal Approximation with XL MIMO Systems: OTA Classification via Trainable Analog Combining
XL-MIMO systems with analog combining perform OTA classification via ELM framework achieving over 90% accuracy with few ms latency under rich fading.
-
What if Pinocchio Were a Reinforcement Learning Agent: A Normative End-to-End Pipeline
The thesis presents Pino, an end-to-end pipeline that supervises reinforcement learning agents with argumentation-based normative advisors, introduces an algorithm for automatic argument extraction, and defines a mitigation strategy for norm avoidance.
-
Non-smooth stochastic gradient descent using smoothing functions
A smoothing stochastic gradient descent algorithm is introduced for non-smooth stochastic compositional optimization, achieving 1/T^{1/4} rate for convex cases and similar guarantees under other convexity settings.
-
Optimizing Multidimensional Scaling in Gini Metric Spaces
Gini MDS replaces Euclidean distance in multidimensional scaling with a rank-and-value-based pseudo-distance controlled by a hyperparameter, claimed to yield more robust embeddings on noisy data than standard MDS.
-
Accelerating Langevin Monte Carlo Sampling: A Large Deviations Analysis
A unified large deviations analysis is proposed to study acceleration mechanisms in variants of overdamped Langevin Monte Carlo methods, supported by numerical experiments.