The Higher Institute for Applied Sciences and Technology (HIAST) (Arabic: الْمَعْهَدُ الْعَالِي لِلْعُلُومِ التَّطْبِيقِيَّةِ وَالتِّكْنُولُوجِيِّ, romanized: al-Maʿhad al-ʿĀlī li l-ʿUlūm at-Taṭbīqīyat wa t-Tiknūlūjī) is center of Excellence for Higher Education, Research & Development in Damascus, Syria. It belongs to the Syrian Scientific Studies and Research Center (SSRC).
إن النظم الحرارية واسعة الانتشار بسبب الكلفة المنخفضة و وثوقيتها العالية وذلك عائد إلى تطوير الحساسات المصفوفية غير المبردة التي أدت إلى الاستغناء عن المسح الميكانيكي وعن التبريد. ولكن تعانيهذه النظم من الانزياح الحراريللمحرق، حيث تتغير قرينة انكسار مواد العدسات البصرية مع تغير درجة الحرارة فيتغير موقع الصورة عن الحساس مؤدياً إلى تشوه الصورة.تقدم هذه المقالة تصميماً بصرياً لنظام حراري بعدستين مصحح حرارياً خفيف الوزن وصغير الحجم. وحسب أفضل معلوماتنا، ليست هذه الحسابات موجودة في أي مرجع.
This paper presents a comprehensive literature survey on image captioning, covering research published between 2018 and 2025. It introduces a novel taxonomy to classify existing approaches into nine major categories, including attention-based models, transformer-based architectures, reinforcement learning, and Vision-Language Pretraining (VLP). A total of 174 peer-reviewed studies are systematically reviewed, with comparative insights drawn across different model architectures, encoding strategies, and learning paradigms. The survey also explores major benchmark datasets such as MS COCO, Flickr30K, and Conceptual Captions, along with evaluation metrics like BLEU, CIDEr, METEOR, ROUGE, SPICE, CHAIR, CLIPScore, and BERTScore. In contrast to prior surveys, this work offers a detailed comparative analysis of state-of-the-art captioning models, highlighting their strengths, limitations, and real-world applicability. Recent models such as PaLI, OSCAR, BLIP-2, and OFA are critically examined in the context of caption quality, generalization, and multimodal alignment. A key research challenge identified across methods is the persistent problem of visual-semantic hallucination, which undermines factual alignment between image content and generated captions. This survey serves as a valuable resource for both newcomers and advanced researchers by offering a structured synthesis of recent developments, challenges, and future directions in the field of image captioning.
The rapid spread of ChatGPT in higher education has created growing pressure on institutions to respond to generative AI in ways that are both educationally meaningful and institutionally responsible. Although existing research increasingly shows that ChatGPT can support student learning and academic performance, these benefits are not automatic and depend on how the tool is used and guided. At the same time, current literature on ChatGPT integration remains fragmented, with most studies focusing separately on pedagogical adaptation, policy and governance, or technology adoption. This study addresses this gap by proposing a conceptual 5D framework for ChatGPT integration in higher education. The framework was developed through literature synthesis and theoretical grounding, drawing on UTAUT, TPACK, Bloom’s Taxonomy, TIM, Innovation Resistance Theory, Responsible AI Governance, and the PDCA continuous improvement model. The proposed framework includes five connected dimensions: Readiness, Pedagogical Integration, Change and Resistance, Governance and Policy, and Monitoring and Improvement. The main contribution of the study is to organize these dimensions into one coherent institution-level framework rather than treating them as separate responses. The framework offers a structured way to understand ChatGPT integration as a broader institutional process that requires preparation, educational alignment, policy direction, support during implementation, and ongoing review over time.
Introduction: Unmanned Aerial Vehicle (UAV) communication networks can move in three-dimensional (3D) space, which can lead to a higher Line of Sight (LoS) probability for wireless communication channels between network nodes. Non-Orthogonal Multiple Access (NOMA) techniques are one of the proposed methods to overcome emerging challenges and meet requirements in the fifth (5G) and sixth (6G) generations of cellular systems. Due to the utilisation of an imperfect Successive Interference Cancellation (SIC) receiver, the performance of the NOMA scheme is vulnerable to error propagation. This paper aims to minimise the power consumption of the NOMA-UAV uplink network, which has limited energy resources. In addition, the impact of Channel State Information (CSI) on resource allocation methods is analysed. Methods: The hybrid multiple access schemes are introduced to minimise the sum power in the NOMA-UAV uplink network, subject to minimum data rate and available resource constraints. The frequency sub-bands and transmission power are jointly optimised to decrease the impact of error propagation in the SIC receiver via three hybrid multiple access models, including (i) a hybrid NOMA and OMA scheme with user grouping on multi-NOMA sub-bands, (ii) a hybrid NOMA and OMA scheme with one NOMA sub-band, and (iii) a multi-NOMA sub-bands scheme. The iterative Sequential Quadratic Programming (SQP) algorithm is employed to solve the optimisation problems. Results: The numerical results show the effectiveness of the proposed hybrid schemes in terms of power consumption compared to the conventional NOMA scheme as benchmarks, where the improvement ratio is more than 20%. The hybrid schemes outperform the NOMA schemes reported in other papers, where the performance was compared relative to the OMA scheme. Discussion: The high diversity of available resources in the hybrid multiple access schemes contributed to achieving the best performance in terms of sum power, which negatively affects the time cost performance. Conclusion: Considering an imperfect SIC receiver, available resources in the NOMA-UAV uplink network are optimised at the ground base station to decrease the transmission power of the UAV communication system, which enhances the energy efficiency of UAVs.
Natural Language Understanding (NLU) plays a crucial role in various applications, yet its performance suffers from weaknesses in handling the complexities of human languages, ranging from lexical ambiguity to high-level reasoning difficulties. Analyzing errors across diverse linguistic phenomena is crucial for NLU improvement, as it will help humans get insights to comprehensively assess models' limitations and capabilities, so optimizing models' generalization. Notably, several benchmarks contain diagnostics datasets designed for investigation and fine-grained error analysis. When highlighting the gaps in the state-of-the-art, we noted that there is no naming convention for macro and micro categories or even a standard set of linguistic phenomena that should be covered. To overcome this gap, we propose an initial hierarchy for Cross-Lingual NLU error analysis. Moreover, we propose a methodology to create an NLI hierarchical framework and applied a case study on Arabic NLU. Moreover, this paper introduces E-CONAN diagnostics dataset, a freely available dataset manually-annotated with coarse-grained and fine-grained categories based on our proposed Arabic hierarchy. E-CONAN dataset helps NLU designers better understand their models by doing error analysis and in-depth investigation. We used E-CONAN to investigate the performance of 9 pretrained language models and 5 LLMs. Results indicate that LLMs outperform pretrained models in world knowledge and commonsense reasoning macro-category, and underperform pretrained models in syntactic macro-category. Moreover, the hardest phenomena for all models is Reasoning, and the easiest phenomena for all pretrained models is Syntactic, and the easiest for LLMs is Lexico-Syntactic.