ECAI 2020

212 papers
The 2020 edition of the European Conference on Artificial Intelligence (ECAI) featured 292 papers, with a strong emphasis on neuro-symbolic integration and the formal verification of neural networks. Research themes prominently included the development of architectures that decouple perception from reasoning and the scaling of safety certificates for uncertain dynamical systems. Significant attention was also directed toward interpretability, specifically bridging the gap between algorithmic outputs and human understanding while addressing the performance costs associated with interpretable models. Technical contributions spanned a variety of domains, including adversarial domain adaptation, tensor-based computation, and link prediction in multiplex networks. Applied research was also represented, such as the use of mobile phone records to estimate malaria parasite mobility in Mozambique. Prolific authors contributing to the proceedings included Meng Wang, Yang Wang, Zhao Zhang, and Francesco Belardinelli. These works collectively highlighted a trend toward making machine learning systems more verifiable, explainable, and sample-efficient through the combination of neural and symbolic methodologies.

Top Research Directions

Research directions most represented by ECAI 2020 papers.

Generalizing Machine Learning Frameworks Across Network Architectures 2
Robust Feature Alignment for Unsupervised Domain Adaptation 2
Decoupling Search from Storage in Multi-Objective Optimization 1
Learned Clause Retention in Incremental Constraint Solving 1
Hierarchical Feature Specialization in Deep Networks 1
Rational Verification of Multi Agent Systems 1
Diagnostic Trace Generation in Formal Verification 1
Enforcing Bidirectional Consistency in Latent Spaces 1
Object Counting via Spatial Density Estimation 1
Pinpointing Minimal Conflicts in Overconstrained Systems 1
Unsupervised Discovery of Latent Behavioral Personas 1
Relational Invariance for Robust Domain Adaptation 1
Data Efficiency Through Neuro Symbolic Integration 1
Solving Word Equations with Length Constraints 1
Tracking Dialogue Progress with Discrete State Transitions 1

Top Authors

AuthorPapers
Jose Hernandez-Orallo4
Fernando Martínez-Plumed3
Francesco Belardinelli3
Frank Neumann3
Meng Wang3
Nuria M Oliver3
Thomas Eiter3
Zhao Zhang3
Alessandro Sperduti2
Aneta Neumann2
Claudia Plant2
Deyi Xiong2
Fausto Giunchiglia2
Giuseppe Marra2
Hisao Ishibuchi2
Jingyu Wang2
Jiuyong Li2
Jong-Seok Lee2
Jörg Rothe2
Jun Zhou2

Impactful Papers

Papers from ECAI 2020 that shaped research clusters.

Foundational

Influential

Papers

Title
Understanding Individual Behaviour: From Virtual to Physical Patterns
Uniform State Abstraction for Reinforcement Learning
Unsupervised Adversarial Learning of Anomaly Detection in the Wild
Unsupervised Learning of Interpretable Dialog Models
Variational Objectives for Markovian Dynamics with Backward Simulation
Verification of Neural Networks: Enhancing Scalability Through Pruning
Verification of Recurrent Neural Networks for Cognitive Tasks via Reachability Analysis
Verifying Strategic Abilities in Multi-Agent Systems via First-Order Entailment
VT-LINKER: Visual-Textual-Knowledge Entity Linker
Weak Supervision Helps Emergence of Word-Object Alignment and Improves Vision-Language Tasks
Weighted LARS for Quantitative Stream Reasoning
What Makes a Better Companion? Towards Social & Engaging Peer Learning

« prev Page 5/5