The two-volume set CCIS 2171 and 2172 constitutes the refereed best papers from the 5th International Conference on Deep Learning Theory and Applications, DeLTA 2024, which took place in Dijon, France, during July 10-11, 2024. The 44 papers included in these proceedings were carefully reviewed and selected from a total of 70 submissions.

Deep Learning for Wearable Biometrics.- A Deep Learning-Based Plant Disease Detection and Classification for Arabica Coffee Leaves.- CNN-N-BEATS: Novel Hybrid Model for Time-Series Forecasting.- Development and Applications of Gesture-Controlled Drones: Advances in Hand Gesture Recognition for Aerial Navigation.- Towards Natural-Sounding Speech to Text in English.- Scoping Review of Active Learning Strategies and Their Evaluation Environments for Entity Recognition Tasks.- Solar Activity Impact on Firefighter Interventions: Factors Analysis.- Detecting Flow via a Machine Learning Model in a MOOC Context.- Online Job Posting Authenticity Prediction with Machine and Deep Learning: Performance Comparison Between N-Gram and TF-IDF.- Evolving Deep Architectures: A New Blend of CNNs and Transformers Without Pre-Training Dependencies.-Closing the Sim-to-Real Gap: Enhancing Autonomous Precision Landing of UAVs with Detection-Informed Deep Reinforcement Learning.- Mitigating Class Imbalance in Healthcare AI Image Classification: Evaluating the Efficacy of Existing Generative Adversarial Network.- More than Noise: Assessing the Viscosity of Food Products Based on Sound Emission.- Vector Analysis of Deep Neural Network Training Process.- Secure Coalition Formation for Federated Machine Learning.- Action Conditioned Attention Encoder-Decoder and Discriminator for Human Motion Generation.- Citation Polarity Identification in Scientific Research Articles Using Deep Learning Methods.- Exploring Physiology-Based Classification of Flow During Musical Improvisation in Mixed Reality.- Refining Weights for Enhanced Object Similarity in Multi-Perspective 6D of Pose Estimation and 3D Object Detection.- End-to-End Video Surveillance Framework for Anomaly Detection and Person Re-Identification.- Few-Shot Learning with Novelty Detection.