Volkan Kılıç

Image captioning has become a key task in vision-language research, bridging visual understanding and natural language generation. However, existing image captioning datasets and models are primarily optimized for sighted individuals, often failing to address the needs of visually impaired individuals who rely on captions for scene understanding and navigation. To address this gap, we introduce ENVISAGE, a large-scale image captioning dataset curated with input from visually impaired individuals. It contains 53,452 images with accessibility-focused captions that emphasize spatial relations and object affordances. Beyond the dataset contribution, we present the first framework for conformal uncertainty estimation in image captioning, providing finite-sample coverage guarantees that improve confidence in caption quality. We benchmark both baseline (Show and Tell, Show, Attend and Tell) and state-of-the-art (BLIP and BLIP-2) captioning models on ENVISAGE under zero-shot and fine-tuned settings. Results demonstrate that conformal predictions achieve reliable coverage guarantees across diverse image categories. ENVISAGE and our uncertainty-aware framework establish a robust and accessible foundation for image captioning systems.

Integrating artificial intelligence (AI) into biosensor technology enables data processing, quantitative analysis, real‐time decision‐making, and adaptive sensing capabilities through advanced pattern recognition and predictive modeling. In addition, AI has the potential to drive innovation in the design of advanced materials for biosensing applications by reducing the reliance on trial‐and‐error methods. This review explores the transformative impact of AI on biosensor technology in the context of historical development, current status, and future prospects. It begins with an overview of the evolution of AI, biosensor technology, and their integration. Comparative analysis of AI‐driven innovations in optical, fluorometric, and electrochemical biosensors is presented, highlighting how AI can improve sensor performance. The role of advanced materials on the development of AI‐assisted biosensors is also discussed as the choice of material has a profound effect on biosensor capabilities. Applications of AI‐assisted biosensors are comprehensively explored across healthcare, environmental monitoring, food safety, and agriculture. This study concludes by addressing challenges, opportunities, ethical concerns, and future research directions, providing a comprehensive and up‐to‐date resource for researchers.

Generating natural language descriptions of an image, namely image captioning, has received much attention in computer vision and natural language processing. Recent image captioning models are mainly based on the encoder-decoder framework in which visual information is extracted by an encoder, e.g. using convolutional neural network (CNN), and captions are generated by a decoder, e.g. using recurrent neural network (RNN). Although this framework is promising for image captioning, there are still issues in the RNN decoder for exploiting the visual information to generate grammatically and semantically correct captions. More specifically, the RNN decoder has limited ability in dealing with long-term complex dependencies, leading to ineffective use of contextual information from the encoded data. To address this issue, in this paper, we introduce a multi-layer gated recurrent unit (ML-GRU) within the conventional RNN decoder, which enables the modulation of the relevant information flow inside the unit, and thus leads to the generation of semantically coherent captions. The proposed ML-GRU-based RNN decoder has been extensively evaluated on the MSCOCO dataset, and experimental results demonstrate the advantage of our proposed approach over the state-of-the-art approaches across multiple performance metrics.

Artificial intelligence (AI) and smartphones have attracted significant interest in microfluidic paper‐based colorimetric sensing due to their convenience and robustness. Recently, AI‐based classification of colorimetric assays has been increasingly reported. However, quantitative evaluation remains a challenge, as classification aims to categorize the color change into discrete class labels rather than a quantity. Therefore, in this study, an AI‐based regression model with enhanced accuracy is developed and integrated into a microfluidic paper‐based analytical device for simultaneous colorimetric measurements of glucose, cholesterol, and pH. The model is also embedded into a smartphone via a custom‐designed Android application named ChemiCheck to complete on‐site colorimetric quantification without internet access in under 1 s. The results demonstrate that the integrated system is able to sensitively detect both glucose (limit of detection [LOD]: 131 ) and cholesterol (LOD: 217 ), concluding the entire analysis within minutes while maintaining a maximum root mean square error of 0.386. Overall, the integrated platform holds great promise for point‐of‐care testing and offers numerous advantages, including easy‐to‐use operation, rapid response, low‐cost, high selectivity, and consistent repeatability, particularly in nonlaboratory and resource‐limited environments.

Microneedle arrays are minimally invasive devices that have been extensively investigated for the transdermal/intradermal delivery of drugs/bioactives. Here, we demonstrate the release of bioactive molecules (estradiol, melatonin and meropenem) from poly(2-hydroxyethyl methacrylate), pHEMA, hydrogel-based microneedle patches in vitro. The pHEMA hydrogel microneedles had mechanical properties that were sufficiently robust to penetrate soft tissues (exemplified here by phantom tissues). The bioactive release from the pHEMA hydrogel-based microneedles was fitted to various models (e.g., zero order, first order, second order). Such pHEMA microneedles have potential application in the transdermal delivery of bioactives (exemplified here by estradiol, melatonin and meropenem) for the treatment of various conditions.

The release of metformin, a drug used in the treatment of cancer and diabetes, from poly(2‐hydroxyethyl methacrylate), pHEMA, hydrogel‐based microneedle patches is demonstrated in vitro. Tuning the composition of the pHEMA hydrogels enables preparation of robust microneedle patches with mechanical properties such that they would penetrate skin (insertion force of a single microneedle to be ≈40 N). Swelling experiments conducted at 20, 35, and 60 °C show temperature‐dependent degrees of swelling and diffusion kinetics. Drug release from the pHEMA hydrogel‐based microneedles is fitted to various models (e.g., zero order, first order, second order). Such pHEMA microneedles have potential application for transdermal delivery of metformin for the treatment of aging, cancer, diabetes, etc.