BenchCouncil Proceedings on Evaluation Science and Engineering: Evaluatology (PEvaluation) serves as the official Proceedings of Bench 2026.
PEvaluation is an open-access, continuously published Proceedings series dedicated to Evaluation Science and Engineering. It publishes high-quality research contributions in areas including evaluation theory and methodology, evaluation engineering, benchmarks, datasets, measurement and testing, evaluation standards, and cross-disciplinary evaluation practices.
Topics of interest include, but are not limited to:
-
Evaluation Science and Methodology
Mathematical modeling and formal specification of evaluation requirements
Development and evolution of evaluation models
Evaluation methodology and theoretical foundations
Design and implementation of evaluation systems
Evaluation risk modeling and quantitative analysis
Cost modeling and optimization for evaluations
Accuracy modeling and error propagation analysis
Evaluation traceability
Identification and standardization of evaluation conditions
Equivalent evaluation conditions and their verification
Experimental design methodologies
Statistical analysis techniques for evaluation
Identification and elimination of confounding factors
Analytical modeling and model validation
Simulation and emulation-based modeling and validation
Domain-specific evaluation methodologies
-
Evaluation Engineering
Benchmark design and implementation
Benchmark traceability
Construction of equivalent evaluation conditions
Evaluation metric and index system design
Scale design and standardization
Evaluation standard design and implementation
Evaluation tools and toolchains
Real-world evaluation systems
Evaluation platforms and testbeds
Industrial evaluation practices
-
Datasets
Dataset construction and development
Dataset quality evaluation
Dataset documentation and metadata standards
Dataset collection, validation, and verification protocols
Dataset reproducibility and reuse
Dataset resampling and meta-analysis techniques
Large-scale data generation while preserving data characteristics
Evaluation frameworks for data-generation experiments
Data sharing infrastructures for reproducible research
-
Benchmarking
Benchmark design and construction
Benchmark suites and benchmark ecosystems
Benchmark validation and maintenance
Benchmark evolution methodologies
Leaderboards and ranking systems
State-of-the-art and state-of-the-practice analysis
Industrial benchmarking practices
Real-world application and system evaluation
Evaluation of emerging technologies in practical scenarios
-
Measurement and Testing
Workload characterization
Instrumentation, sampling, tracing, and profiling
Measurement methodologies for large-scale systems
Testing methodologies and frameworks
Measurement-driven knowledge discovery
Performance modeling and bottleneck analysis
Scalability and efficiency evaluation
Monitoring and visualization of measurement data
Reproducible measurement practices
Re-evaluation of previous empirical measurements and conclusions
-
Algorithm Evaluation and Optimization
Benchmark-driven algorithm evaluation
Standardized evaluation of algorithms
Algorithm performance analysis
Generalization, robustness, fairness, and reliability evaluation
Hardware-aware algorithm evaluation and co-design
Accuracy-cost and efficiency-quality trade-off analysis
Adaptive evaluation under dynamic environments
Reproducibility of algorithm evaluation
Optimization driven by evaluation feedback
-
AI and Intelligent System Evaluation
Foundation model evaluation
Large language model evaluation
AI benchmark development
Agent evaluation
Multimodal AI evaluation
AI safety evaluation
Robustness evaluation
Fairness evaluation
Explainability evaluation
Human-centered evaluation
Real-world AI system evaluation
-
Cross-disciplinary Evaluation Applications
Bench 2026 welcomes evaluation research and engineering practices in, but not limited to:
Computer science
Artificial intelligence
Medicine and healthcare
Biology
Education
Finance and economics
Business and management
Psychology
Social sciences
Earth sciences
Environmental sciences
Transportation
Energy
Manufacturing
Digital humanities
Other scientific and engineering disciplines