| 1. |
- Munir, M. Adeel, et al.
(författare)
-
Blockchain Adoption for Sustainable Supply Chain Management : Economic, Environmental, and Social Perspectives
- 2022
-
Ingår i: Frontiers in Energy Research. - : Frontiers Media S.A.. - 2296-598X. ; 10
-
Tidskriftsartikel (refereegranskat)abstract
- Due to the rapid increase in environmental degradation and depletion of natural resources, the focus of researchers is shifted from economic to socio-environmental problems. Blockchain is a disruptive technology that has the potential to restructure the entire supply chain for sustainable practices. Blockchain is a distributed ledger that provides a digital database for recording all the transactions of the supply chain. The main purpose of this research is to explore the literature relevant to blockchain for sustainable supply chain management. The focus of this review is on the sustainability of the blockchain-based supply chain concerning environmental conservation, social equality, and governance effectiveness. Using a systematic literature review, a total of 136 articles were evaluated and categorized according to the triple bottom-line aspects of sustainability. Challenges and barriers during blockchain adoption in different industrial sectors such as aviation, shipping, agriculture and food, manufacturing, automotive, pharmaceutical, and textile industries were critically examined. This study has not only explored the economic, environmental, and social impacts of blockchain but also highlighted the emerging trends in a circular supply chain with current developments of advanced technologies along with their critical success factors. Furthermore, research areas and gaps in the existing research are discussed, and future research directions are suggested. The findings of this study show that blockchain has the potential to revolutionize the entire supply chain from a sustainability perspective. Blockchain will not only improve the economic sustainability of the supply chain through effective traceability, enhanced visibility through information sharing, transparency in processes, and decentralization of the entire structure but also will help in achieving environmental and social sustainability through resource efficiency, accountability, smart contracts, trust development, and fraud prevention. The study will be helpful for managers and practitioners to understand the procedure of blockchain adoption and to increase the probability of its successful implementation to develop a sustainable supply chain network.
|
|
| 2. |
- Munir, Muhammad Akhtar, et al.
(författare)
-
Bridging Precision and Confidence: A Train-Time Loss for Calibrating Object Detection
- 2023
-
Ingår i: 2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR). - : IEEE COMPUTER SOC. - 9798350301298 - 9798350301304 ; , s. 11474-11483
-
Konferensbidrag (refereegranskat)abstract
- Deep neural networks (DNNs) have enabled astounding progress in several vision-based problems. Despite showing high predictive accuracy, recently, several works have revealed that they tend to provide overconfident predictions and thus are poorly calibrated. The majority of the works addressing the miscalibration of DNNs fall under the scope of classification and consider only in-domain predictions. However, there is little to no progress in studying the calibration of DNN-based object detection models, which are central to many vision-based safety-critical applications. In this paper, inspired by the train-time calibration methods, we propose a novel auxiliary loss formulation that explicitly aims to align the class confidence of bounding boxes with the accurateness of predictions (i.e. precision). Since the original formulation of our loss depends on the counts of true positives and false positives in a mini-batch, we develop a differentiable proxy of our loss that can be used during training with other application-specific loss functions. We perform extensive experiments on challenging in-domain and out-domain scenarios with six benchmark datasets including MS-COCO, Cityscapes, Sim10k, and BDD100k. Our results reveal that our train-time loss surpasses strong calibration baselines in reducing calibration error for both in and out-domain scenarios. Our source code and pre-trained models are available at https://github.com/akhtarvision/bpc_calibration
|
|
