Dag DLT

Directed acyclic graph (DAG) techniques are increasingly used in various fields to represent and learn causal relationships or dependencies within data. Current research focuses on developing efficient algorithms for learning DAG structures from data, including score-based approaches for causal Bayesian networks and differentiable probabilistic models for faster and more robust optimization. These advancements are improving the accuracy and scalability of causal discovery, with applications ranging from federated learning (where DAGs enhance privacy and efficiency) to natural language processing (where they enable faster and more controlled text generation). The overall impact is a more robust and versatile toolkit for analyzing complex systems and building more intelligent systems.

Papers

October 31, 2024

$ψ$DAG: Projected Stochastic Approximation Iteration for DAG Structure Learning
Klea Ziu, Slavomír Hanzely, Loka Li, Kun Zhang, Martin Takáč, Dmitry Kamzolov
Stochastic Gradient Descent Gradient Based Directed Acyclic Graph Structure Learning Iterative Stochastic Dag DLT

August 20, 2024

A Full DAG Score-Based Algorithm for Learning Causal Bayesian Networks with Latent Confounders
Christophe Gonzales, Amir-Hosein Valizadeh
Practical Algorithm Probabilistic Graphical Model Latent Confounders Causal Bayesian Network Non Causal Dag DLT

May 21, 2024

Better Simulations for Validating Causal Discovery with the DAG-Adaptation of the Onion Method
Bryan Andrews, Erich Kummerfeld
Causal Discovery Causal Model Causal Structure Learning Simulation Fidelity Dag DLT

April 10, 2024

Control-DAG: Constrained Decoding for Non-Autoregressive Directed Acyclic T5 using Weighted Finite State Automata
Jinghong Chen, Weizhe Lin, Jingbiao Mei, Bill Byrne
Language Generation Neural Machine Translation Constrained Decoding Schema Guided Dialogue Weighted Finite Linear Non Gaussian Acyclic Model Dag DLT

August 25, 2023

DAG-ACFL: Asynchronous Clustered Federated Learning based on DAG-DLT
Xiaofeng Xue, Haokun Mao, Qiong Li
Asynchronous Federated Learning Dag DLT FL Framework

March 4, 2023

DAG Matters! GFlowNets Enhanced Explainer For Graph Neural Networks
Wenqian Li, Yinchuan Li, Zhigang Li, Jianye Hao, Yan Pang
Graph Neural Network Energy Based GFlowNets GNN Explainers Generative Grammar Dag DLT

December 17, 2022

DAG: Depth-Aware Guidance with Denoising Diffusion Probabilistic Models
Gyeongnyeon Kim, Wooseok Jang, Gyuseong Lee, Susung Hong, Junyoung Seo, Seungryong Kim
Diffusion Model Generative Model Denoising Diffusion Probabilistic Model Depth Map Human Guidance Depth Information Depth Aware Feature Dag DLT

October 27, 2022

Learning Discrete Directed Acyclic Graphs via Backpropagation
Andrew J. Wren, Pasquale Minervini, Luca Franceschi, Valentina Zantedeschi
Back Propagation Directed Acyclic Graph Discrete Activation Dag DLT

September 26, 2022

An Energy Optimized Specializing DAG Federated Learning based on Event Triggered Communication
Xiaofeng Xue, Haokun Mao, Qiong Li, Furong Huang
Timely Communication Event Description Directed Acyclic Graph Energy Policy Research Federated Graph Learning Dag DLT Federated Causal

March 16, 2022

Differentiable DAG Sampling
Bertrand Charpentier, Simon Kibler, Stephan Günnemann
Variational Inference Continuous Optimization Dag DLT