Hardness of Samples Need to be Quantified for a Reliable Evaluation System: Exploring Potential Opportunities with a New Task
This addresses the issue of overestimating AI capabilities for researchers and practitioners, though it is incremental as it builds on existing evaluation methods.
The authors tackled the problem of unreliable model evaluation due to unquantified sample hardness, proposing a Data Scoring task to assign difficulty scores to unannotated benchmark samples and showing that models perform better on easier samples.
Evaluation of models on benchmarks is unreliable without knowing the degree of sample hardness; this subsequently overestimates the capability of AI systems and limits their adoption in real world applications. We propose a Data Scoring task that requires assignment of each unannotated sample in a benchmark a score between 0 to 1, where 0 signifies easy and 1 signifies hard. Use of unannotated samples in our task design is inspired from humans who can determine a question difficulty without knowing its correct answer. This also rules out the use of methods involving model based supervision (since they require sample annotations to get trained), eliminating potential biases associated with models in deciding sample difficulty. We propose a method based on Semantic Textual Similarity (STS) for this task; we validate our method by showing that existing models are more accurate with respect to the easier sample-chunks than with respect to the harder sample-chunks. Finally we demonstrate five novel applications.