Introduction. High-frequency proton therapy is a promising breast cancer (BC) treatment. Earlier, we revealed that the extent of nanoparticle (NPs) uptake in BC cells indicates their metastatic potential (MP). The research investigated BC cell nanoparticles absorption, comparing parental cells with their proton irradiation survivors in different environmental stiffnesses.

Materials and Methods. Low-MP MCF7 and high-MP MDA-MB-231 cell lines and their fractions survived after 6Gr-dose proton beams MCF7RP and MDA-MB-231RP, respectively, were seeded onto either rigid plastic flasks or soft fibrin gels for 24 hours, followed by incubation with fluorescent 200 nm nanoparticles and nuclear stain. After washing, cells were imaged in fluorescence channels, and Pearson’s colocalization coefficient was calculated.

Results. On rigid plastic, MDA-MB-231 cells internalized 1.35-fold more nanoparticles than MCF7 cells (p<0.0017). NP uptake further increased 1.6-fold in MDA-MB-231RP cells (p<0.0001), whereas MCF7RP cells showed no change. Overall, MDA-MB-231RP uptake was 1.86-fold higher than MCF7RP (p<0.0001), suggesting increase in MP. On soft fibrin, NP uptake by MDA-MB-231 cells decreased 1.33-fold, while uptake by MCF7 cells increased 1.18-fold compared with rigid plastic, resulting in comparable uptake levels between the two cell-lines. In contrast, MCF7RP cells showed reduced NP uptake, while MDA-MB-231RP cells were 1.47-fold more efficient (p<0.0172).

Conclusions. Proton irradiation BC cell survivors exhibited a pattern of nanoparticle uptake that correlates with their MP and the stiffness of the microenvironment. Hence, combining proton therapy with methods to soften the tumor microenvironment could help reduce the risk of metastasis after treatment.

Introducion: This research explored the incorporation of Malva sylvestris L. (common mallow) extracts into yogurt to develop a functional food with enhanced nutritional value. The study focused on evaluating the plant's antioxidant and antibacterial activities to assess its potential for creating health-promoting dietary products.

Methods: The methodology involved extracting bioactive compounds from Malva sylvestris using maceration. The resultant extract underwent detailed phytochemical analysis to quantify total phenolics, flavonoids, tannins, and sugars. Its antioxidant efficacy was thoroughly evaluated using four distinct assays: DPPH, ABTS, FRAP, and Phenanthroline. The extract was subsequently fortified into yogurt, and the resulting product was rigorously analyzed for its physicochemical properties, sensory attributes, and antimicrobial effects.

Results: The findings demonstrated that Malva sylvestris is abundant in bioactive constituents, notably polyphenols (217 µg GAE/ml) and sugars (131.59 µg GE/ml). This rich phytochemical profile was directly linked to a potent antioxidant capacity, confirmed by remarkably low IC₅₀ values in the radical scavenging assays. Fortification of yogurt with the extract did not significantly alter its overall sensory or physical characteristics, confirming compatibility with the food matrix. The lactic acid bacterial count in the final product complied with the Codex Alimentarius standard of <10⁷ CFU/ml. A dose-dependent increase in pH was observed with higher extract concentrations, which slightly influenced the product's properties.

Conclusion: Our findings confirm Malva sylvestris L. as a potent natural antioxidant source, whose successful use in yogurt reveals significant potential for functional food and nutraceutical development.

Drought constitutes a major threat to food security and the sustainability of agricultural systems in tropical developing countries. These systems are highly vulnerable due to their strong dependence on rainfall, their low resilience to climatic hazards, and the scarcity of reliable data. This study proposes a methodological framework for assessing and forecasting agricultural drought exposure in semi-arid tropical regions, using Haute Matsiatra in Madagascar as a case study. The framework integrates multi-source geospatial data with artificial intelligence to construct and predict a drought exposure index. Climate data are retrieved from the NASA POWER LARC platform and local institutions, while Landsat satellite imagery (30 m resolution) is sourced from the USGS platform to derive radiometric indices. The approach consists of three main steps: (i) selection and preprocessing of climate variables, satellite imagery, and radiometric indices such as NDWI and SPEI; (ii) development of a composite exposure index through weighted aggregation, incorporating local expertise using Saaty’s Analytic Hierarchy Process (AHP); and (iii) forecasting the index with a Long Short-Term Memory (LSTM) recurrent neural network, using horizons of 6 to 36 months aligned with the local agricultural calendar. To adapt to these horizons, daily data are aggregated to reduce noise due to data granularity. The outputs will be presented as thematic maps illustrating exposure levels under different horizons and climate scenarios. This framework offers a decision-support tool for climate risk management and can be replicated or adapted to other regions facing similar vulnerabilities.

The rapid spread of misinformation on social media has emerged as a major societal issue. Over 40% of British social media news‑sharers admitted they had shared inaccurate or fake news. The extensive distribution of false information causes public trust deterioration while modifying public opinions and potentially destabilizing social and political systems. There are profound challenges due to this hard to detect, hard to stop reality and the financials and sociatal implications are remarkable. As an attempt to limit the challenges created from misinformation this paper introduces some preliminary work on detection of fake news and verification of their reliability based on online content. Large language models (LLMs) are being used along with natural language processing (NLP) techniques to evaluate news articles through their linguistic and contextual characteristics. Several models are compared on how they can typically identify typical indicators of misinformation through the analysis of extensive verified datasets to develop an ability to classify content as authentic or fabricated . This work  has been through thorough testing to determine its operational effectiveness and dependability after completion. We present a relatively easy-to-use tool which enables a wide range of people also for them without a background computer science to  easily verify news accuracy before sharing or trusting it. This work could help to stop false information from spreading while promoting fact based discussions and improving digital literacy skills. The research demonstrates how technology fights the fake news crisis to create an informed digital environment which supports public conversation protection and information integrity in the modern digital age.

One of the major challenges in the Internet of Things (IoT) is processing complex data. Data sourced from numerous sensors, cameras, and network logs are continuously evolving and are also multimodal. Having multiple data sources and a dynamic environment gives rise to concept drift. The precision and reliability of Static Machine Learning Models are impacted by concept drift. This is where our framework, Adaptive Multimodal Long Short-Term Memory (AM-LSTM), comes in. It uses a unique technique that combines online learning with smart fusion to ensure accurate and robust implementation in a dynamic system. The system learns relentlessly, adapting to all inputs, and constantly ensures the data stays relevant. To implement this, the following methodologies were adopted: First, dedicated LSTMs help isolate and learn the temporal patterns in each data modality. Secondly, the attention-based fusion mechanism dynamically selects the most appropriate information across these modalities and is tolerant to missing data issues. Thirdly, the concept drift was addressed through the window technique. The method used makes ongoing assessments of the prediction errors, and when a substantial change is observed, a relevant retraining cycle is detected. The AM-LSTM model was assessed using UNSW-NB15 and Edge-IoT dataset benchmarks and was found to function effectively. The model yielded a score of 88.7% and an F1 score of 0.85. It was reactive to concept drift by adapting to changes after just 620 samples, which outperformed all benchmark models. The 47 milliseconds delay in every batch update indicates the performance and robustness of real-time IoT systems.

This study investigated the effects of germination and toasting on the physicochemical properties of protein concentrate extracted from ‘guna’ (Cucumis melon Linn) kernel. Conducted at the Department of Animal Production, Federal University of Technology, Minna, the experiment involved measuring proximate composition, physicochemical properties, and amino acid profiles, with results expressed as means of three trials ± standard deviation. The proximate analysis revealed that the untreated ‘guna’ kernel (CGK) had a crude fat content of 2.56% ± 0.01. In comparison, germinated ‘guna’ kernel (GGK) showed a moisture content of 4.13% ± 0.09. Crude fiber was 3.33% ± 0.01 in CGK, and crude protein was highest in both GGK and toasted ‘guna’ kernel (TGK) at 77.55% ± 0.04. Ash content was 2.47% ± 0.04 in CGK and TGK, while carbohydrate content was 12.17% ± 0.02 in TGK. Physicochemical parameters such as titrable acidity, total solids, yield, and dispersibility were recorded, with values including 5.30 ± 0.01 (GGK), 97.20 ± 0.76 (TGK), and 44.93 ± 0.07 (CGK), respectively. Amino acid profiling indicated that the protein concentrate contains all essential amino acids, with notable levels of isoleucine, threonine, lysine, histidine, and methionine in TGK and GGK samples. Statistical analysis using one-way ANOVA confirmed that both toasting and germination significantly improved protein content by 1.84% and 2.06%, respectively. The study recommends further research focusing on protein isolates to explore the effects of these treatments on functional properties, aiming to enhance industrial applications of ‘guna’ kernel protein concentrates.

Human Activity Recognition (HAR) is the keystone for healthcare, smart homes, and mobile computing; nevertheless, the process of putting deep learning models right on edge devices has always been a hard nut to crack because of the high computational requirements. This research article is a perfect example of a work that solves what we might call the paradox of real-time and efficient HAR with accuracy kept intact. In this paper, this is achieved by creating a lightweight convolutional neural network (CNN) and using TensorFlow Lite to perform the quantization and related optimizations on the UCI HAR dataset.

The design reaches 93.5% of the accuracy before quantization. After the quantization, there is a loss of less than 1% in accuracy. The dimension of the model decreases from ~4.2 MB to ~0.9 MB, and the inference latency is almost half, thus making it IoT as well as mobile devices user-friendly. The findings validate the possibility of employing compact CNNs as they can strike a balance between the accuracy of the solution and computational efficiency, thus making it possible to perform HAR on platforms with limited resources.

This is a project that merges the possibility of achieving the highest accuracy for HAR and the feasibility of deploying it in edge devices. The research works that lie ahead include multimodal HAR, lightweight transformer architectures, and real-world streaming applications.

The conventional approaches to sentiment analysis often use only one type of modality, e.g., text or speech; hence, they cannot be used to identify the richness of human expression of emotions. As the concept of deep learning rapidly expands, Multimodal Sentiment Analysis (MSA) has become a potent technology that allows the merging of various sources of data to enhance emotional comprehension. This paper introduces a transformer-based architecture, integrating speech, text and faces to improve the accuracy of sentiment classification. Transformer networks are able to use the self-attention mechanism to capture long-range interactions as well as cross-modal interactions, which are difficult to capture using more traditional recurrent or convolutional models. The suggested system derives textual embeddings based on a pre-trained language model, acoustic features based on spectrogram-based encoders, and visual interpreters based on facial landmark and expression recognition systems. A cross-modal attention fusion approach synchronizes and dynamically balances features between modalities to produce higher-level and more context-initiative sentiment cues. Results of experiments on benchmark datasets, including CMU-MOSEI and IEMOCAP, show that the proposed model provides an accuracy of 87.6 and an F1-score of 86.9, surpassing unimodal and early-fusion baselines by 6.4 and 5.8, respectively. The architecture has been found to be accurate at recognizing subtle or ambiguous emotions. These results show the possibilities of using transformer-based MSA systems in real-life scenarios, such as human–computer interaction, healthcare, social robotics, and in digital learning environments, creating a path to emotionally intelligent and responsive AI systems.

Deep Learning methodologies have been shown to assist vegetation plantations and their stakeholders. Extraction of horticultural plantations from satellite imagery is an important example for agricultural monitoring, food security, and sustainable land management. Traditional techniques often fall short in generalizing across seasons or heterogeneous landscapes, necessitating more adaptive approaches. To address this, we utilized Conditional Generative Adversarial Networks to segment sample plantations of pomegranate and date palm in the arid regions of Rajasthan, India (Jaisalmer and Barmer districts), using Very High-Resolution Satellite (VHRS) imagery from CARTOSAT-2E. The framework, inspired by the existing Pix2Pix image translation model by Philip Isola, incorporated a U-Net generator and Patch-GAN discriminator for segmentation, trained on over 13,000 annotated image–mask pairs. Using loss functions, generator skip connections, and various activation functions, the Pix2Pix U-Net cGAN model demonstrated strong segmentation capabilities. Dice coefficient of 99.49%, IoU of 98.99%, pixel accuracy of 99.28%, precision of 99.66%, and recall of 99.32%. Focal and edge-aware loss functions further enhanced class differentiation, yielding a Dice coefficient of 94.35%, an IoU of 91.52%, a pixel accuracy of 97.95%, a precision of 97.06%, and a recall of 94.09%. While these metrics may have shown a strong performance, the true effectiveness of this approach was demonstrated through accurate vegetation classification and recreating the vegetation masks for extraction and segmentation. The workflow automates plantation detection from VHRS data. It establishes a foundation for future opportunities in horticulture segmentation and analysis, as well as its potential integration into decision-support systems for sustainable resource management.

Background
Taraxacum officinale (common dandelion) is a perennial herb widely valued for its nutritional and medicinal properties having strong potential in food technology as a natural source of bioactive compounds. Its vitamin- and mineral-rich leaves, inulin- and antioxidant-containing roots, and flavonoid-rich flowers highlight its potential for use in functional beverages, dietary supplements, and natural food additives.

Methods
The research included morphological and anatomical analysis, chemical composition determination (polyphenols, flavonoids, tannins, sugars), and the assessment of antioxidant activity using DPPH method. The antibacterial activity of the prepared extracts was assessed using the disk diffusion method on Mueller-Hinton agar. The tested bacterial strains included Gram-negative (Escherichia coli, Pseudomonas aeruginosa) and Gram-positive (Staphylococcus aureus , Bacillus subtilis ) species. This screening aimed to determine the inhibitory potential of the extracts against clinically relevant microorganisms.

Results
Anatomical studies revealed well-defined vascular structures and storage parenchyma, consistent with its taxonomic classification. Phytochemical screening showed high concentrations of polyphenols (33.76 μg GAE/mg), tannins and sugars but lower flavonoids. The extract exhibited strong antioxidant capacity (IC₅₀: 143.97 μg/mL) in DPPH assays, confirming its efficacy against oxidative stress. However, antibacterial tests against standard strains revealed negligible inhibition zones. The results highlight T. officinale's significant antioxidant properties while suggesting limited direct antimicrobial activity, supporting its traditional use in antioxidant-related therapies.

Conclusion
Overall, T. officinale represents a promising candidate for the development of functional foods and nutraceuticals, particularly as a natural antioxidant source. Its incorporation into beverages, supplements, or as an additive could improve both the nutritional value and health-promoting properties of food products.