Large Language Models (LLMs) hold rich implicit knowledge and powerful transferability. In this paper, we explore the combination of LLMs with the human skeleton to perform action classification and description. However, when treating LLM as a recognizer, two questions arise: 1) How can LLMs understand the skeleton? 2) How can LLMs distinguish among actions? To address these problems, we introduce a novel paradigm named learning Skeleton representation with visual-motion knowledge for Action Recognition (SUGAR). In our pipeline, we first utilize off-the-shelf large-scale video models as a knowledge base to generate visual, motion information related to actions. Then, we propose to supervise skeleton learning through this prior knowledge to yield discrete representations. Finally, we use the LLM with untouched pre-training weights to understand these representations and generate the desired action targets and descriptions. Notably, we present a Temporal Query Projection (TQP) module to continuously model the skeleton signals with long sequences. Experiments on several skeleton-based action classification benchmarks demonstrate the efficacy of our SUGAR. Moreover, experiments on zero-shot scenarios show that SUGAR is more versatile than linear-based methods.

Abstract In recent years, large language models (LLMs) have gained significant traction across various domains, including education. This paper explores the application of LLMs in grading programming assignments. By leveraging data collected from existing programming assignments and their corresponding grades, we aim to develop a robust LLM-based grading system. We also incorporate augmented data representing various grading scenarios to enhance the model’s performance and ensure comprehensive coverage across all grading levels. Our approach involves training the LLM on this combined dataset to enable accurate and consistent evaluation of programming assignments. The proposed model, BeGrading, aims to reduce the grading burden on educators and provide timely and objective feedback to students. Compared to the Codestral model, our proposed model demonstrates an absolute difference rate of 19%, equivalent to $$\pm 0.95$$ ± 0.95 out of 5. This is acceptable for using a small, fine-tuned model with optimized data. Additionally, the Codestral model compared to the dataset optimized score shows a difference of 15% equivalent to a margin of $$\pm 0.75$$ ± 0.75 out of 5. Preliminary results demonstrate the potential of LLMs to perform grading tasks with a high degree of reliability, opening avenues for further research and practical applications in automated education systems.

Abstract A mobile ad hoc network (MANET) is a self‐organized and adaptive wireless network formed by dynamically gathering mobile nodes. Since the topology of the network is constantly changing, the issue of routing packets and energy conservation become challenging tasks. In this paper, we propose a cross‐layer design that jointly considers routing and topology control taking mobility and interference into account for MANETs. We called the proposed protocol as Mobility‐aware Routing and Interference‐aware Topology control (MRIT) protocol. The main objective of the proposed protocol is to increase the network lifetime, reduce energy consumption, and find stable end‐to‐end routes for MANETs. We evaluate the performance of the proposed protocol by comprehensively simulating a set of random MANET environments. The results show that the proposed protocol reduces energy consumption rate, end‐to‐end delay, interference while preserving throughput and network connectivity. Copyright © 2010 John Wiley & Sons, Ltd.

A wireless ad hoc network (WANET) is a collection of wireless terminals that communicate with each other without predetermined topology. Since WANET devices are power-limited, network protocols should be designed to prolong the battery lifetime of these devices. In this paper, we propose a cross-layer integration approach for power efficient routing protocol. The proposed cross-layer integration between power control in link layer and routing protocol in network layer aims to maximize the network lifetime. We implement our proposed protocol as an extension to AODV routing protocol. We evaluated the proposed protocol by comprehensively simulating a set of random WANET environments. We simulated six different metrics comparing our proposed protocol with AODV protocol. The results showed that the proposed protocol maximizes the network lifetime, reduces the end-to-end delay, and saves the total energy consumption while achieving the throughput requirement.