Synthesis 2022
111 papers
Synthesis 2022 featured 125 papers covering diverse computational topics, with a significant emphasis on acoustic simulation and domain adaptation. Research in sound propagation explored hybrid wave-geometric methods and neural acoustic fields to achieve realistic spatial audio. Parallel to this, a substantial body of work addressed challenges in machine learning robustness, specifically focusing on dataset bias, domain shift, and unsupervised domain adaptation through entropy minimization and adversarial alignment.
The edition also highlighted advancements in graph-based learning, including applications for fashion compatibility modeling and personalized online content management. Notable contributions explored fundamental graph neural networks and network connectivity optimization. Prolific contributors to the 2022 proceedings included researchers such as Rishiraj Saha Roy, Aidan Hogan, and Axel Polleres. The collective work reflects a trend toward integrating physics-based synthesis with neural methods and refining model performance across heterogeneous data environments through techniques like domain-specific batch normalization.
Top Research Directions
Research directions most represented by Synthesis 2022 papers.
Top Authors
Impactful Papers
Papers from Synthesis 2022 that shaped research clusters.
Influential
Visual Domain Adaptation in the Deep Learning Era
influential
Preventing Negative Transfer from Outlie
Papers
« prev Page 3/3