Access to powerful computational resources during the 2010s fueled a reemergence in research within the field of deep learning (DL) which is a subset of machine learning. DL comprising a convolutional neural network (CNN) has emerged as a transformative and highly accurate method for both image classification and object detection. It has enabled real-time identification and remediation of costly weeds within both row-crop and specialty crop farming systems. Further, advancements in computing technology have reduced the physical hardware size required for model training. This chapter will discuss (1) the importance of site-specific weed management (SSWM), (2) weed identification and detection research, (3) the importance of transfer learning, (4) different weed identification and detection methods, and (5) other relevant advanced weed management technologies. Two case studies that utilized DL for UAS-based and handheld weed identification will be discussed. Specifically, materials needed to conduct these studies, design of experiments, data acquisition procedures, preprocessing of data, training DL models, and their results are discussed. Results of the UAS-based weed detection study yielded a 91.48% and 86.13% average precision (AP) at an Intersection over Union (IoU) of 0.25, for monocot and dicot weeds, respectively. Image classification models that were trained to identify the four weeds present within the 4Weeds dataset yielded 43.28%, 26.30%, 89.89%, and 57.80% AP for common weeds, namely, cocklebur, foxtail, redroot pigweed, and giant ragweed, respectively.