Recently, several methods have been proposed to explain the predictions of recurrent neural networks (RNNs), in particular of LSTMs. The goal of these methods is to understand the network's decisions by assigning to each input variable, e.g., a word, a relevance indicating to which extent it contributed to a particular prediction. In previous works, some of these methods were not yet compared to one another, or were evaluated only qualitatively. We close this gap by systematically and quantitatively comparing these methods in different settings, namely (1) a toy arithmetic task which we use as a sanity check, (2) a five-class sentiment prediction of movie reviews, and besides (3) we explore the usefulness of word relevances to build sentence-level representations. Lastly, using the method that performed best in our experiments, we show how specific linguistic phenomena such as the negation in sentiment analysis reflect in terms of relevance patterns, and how the relevance visualization can help to understand the misclassification of individual samples.
PurposeCold supply chain (CSC) distribution systems are vital in preserving the integrity and freshness of transported temperature sensitive products. CSC is also known to be energy intensive with a significant emission footprint. As a result, CSC requires strict monitoring and control management system during storage and transportation to improve safety and reduce profit losses. In this research, a systematic review of recent literature related to the distribution of food CSC products is presented and possible areas to extend research in modeling and decision-making are identified.Design/methodology/approachThe paper analyzes the content of 65 recent articles related to CSC and perishable foods. Several relevant keywords were used in the initial search, which generated a list of 214 articles. The articles were screened based on content relevance in terms of food vehicle routing modeling and quality. Selected articles were categorized and analyzed based on cost elements, modeling framework and solution approach. Finally, recommendations for future research are suggested.FindingsThe review identified several research gaps in CSC logistics literature, where more focused research is warranted. First, the review suggests that dynamic vehicle modeling and routing while considering products quality and environmental impacts is still an open area for research. Second, there is no consensus among researchers in terms of quality degradation models used to assess the freshness of transported cold food. As a result, an investigation of critical parameters and quality modeling is warranted. Third, and due to the problem complexity, there is a need for developing heuristics and metaheuristics to solve such models. Finally, there is a need for extending the single product single compartment CSC to multi-compartment multi-temperature routing modeling.Originality/valueThe article identified possible areas to extend research in CSC distribution modeling and decision-making. Modified models that reflect real applications will help practitioners, food authorities and researchers make timely and more accurate decisions that will reduce food waste and improve the freshness of transported food.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.