Recent studies on plant disease diagnosis using machine learning (ML) have highlighted concerns about the overestimated diagnostic performance due to inappropriate data partitioning, where training and test datasets are derived from the same source (domain). Plant disease diagnosis presents a challenging classification task, characterized by its fine-grained nature, vague symptoms, and the extensive variability of image features within each domain. In this study, we propose the concept of Discriminative Difficulty Distance (DDD), a novel metric designed to quantify the domain gap between training and test datasets while assessing the classification difficulty of test data. DDD provides a valuable tool for identifying insufficient diversity in training data, thus supporting the development of more diverse and robust datasets. We investigated multiple image encoders trained on different datasets and examined whether the distances between datasets, measured using low-dimensional representations generated by the encoders, are suitable as a DDD metric. The study utilized 244,063 plant disease images spanning four crops and 34 disease classes collected from 27 domains. As a result, we demonstrated that even if the test images are from different crops or diseases than those used to train the encoder, incorporating them allows the construction of a distance measure for a dataset that strongly correlates with the difficulty of diagnosis indicated by the disease classifier developed independently. Compared to the base encoder, pre-trained only on ImageNet21K, the correlation higher by 0.106 to 0.485, reaching a maximum of 0.909.

Numerous studies have explored image-based automated systems for plant disease diagnosis, demonstrating impressive diagnostic capabilities. However, recent large-scale analyses have revealed a critical limitation: that the diagnostic capability suffers significantly when validated on images captured in environments (domains) differing from those used during training. This shortfall stems from the inherently limited dataset size and the diverse manifestation of disease symptoms, combined with substantial variations in cultivation environments and imaging conditions, such as equipment and composition. These factors lead to insufficient variety in training data, ultimately constraining the system's robustness and generalization. To address these challenges, we propose Few-shot Metric Domain Adaptation (FMDA), a flexible and effective approach for enhancing diagnostic accuracy in practical systems, even when only limited target data is available. FMDA reduces domain discrepancies by introducing a constraint to the diagnostic model that minimizes the "distance" between feature spaces of source (training) data and target data with limited samples. FMDA is computationally efficient, requiring only basic feature distance calculations and backpropagation, and can be seamlessly integrated into any machine learning (ML) pipeline. In large-scale experiments, involving 223,015 leaf images across 20 fields and 3 crop species, FMDA achieved F1 score improvements of 11.1 to 29.3 points compared to cases without target data, using only 10 images per disease from the target domain. Moreover, FMDA consistently outperformed fine-tuning methods utilizing the same data, with an average improvement of 8.5 points.

Learning-based solutions are efficient, but are they trustworthy enough to be embedded in a robot cooperating with or assisting humans? In this blogpost, Daniele Meli explores this question, and reviews logic programming as a route to trustworthy autonomous (and cooperative) robotic systems. As part of the 37th AAAI Conference on Artificial Intelligence (AAAI2023), 32 different workshops were held, covering a wide range of topics. We heard from the organisers of four of these workshops, who told us their key takeaways from their respective events. These were split into two articles: 1) #AAAI2023 workshops round-up 1: AI for credible elections, and responsible human-centric AI, and 2) #AAAI2023 workshops round-up 2: health intelligence and privacy-preserving AI. Hosted by the Alan Turing Institute, AI UK is a two-day conference that showcases artificial intelligence and data science research, development, and policy in the UK. This year, the event took place on 21 and 22 March, and we covered the panel discussion session on the role and impact of science journalism. AAAI have updated their publication policy to deal with AI systems: "It is AAAI's policy that any AI system, including Generative Models such as Chat-GPT, BARD, and DALL-E, does not satisfy the criteria for authorship of papers published by AAAI and, as such, also cannot be used as a citable source in papers published by AAAI".