LNM Institute of Information Technology

Current AI models process input through a single lens. The human brain never does this—it cross-references every sense against every other sense, flags conflicts, and only calls on expensive conscious reasoning when something doesn't add up. We present a complete architecture for an artificial brain that mirrors this design: specialized Small Language Models (SLMs) as parallel sensory cortices, a lightweight conflict detector as the anterior cingulate cortex, an expensive reasoning model as the prefrontal cortex activated only on demand, a streaming identity core as the default mode network, episodic memory via vector databases, slow knowledge consolidation via sleep-cycle fine-tuning, and a neuromodulator reward system that shapes all behavior over time. The system is born when started, develops personality through experience, sleeps to consolidate, and dies when stopped. Every component maps to a specific brain structure. Every design decision is grounded in neuroscience research. The architecture is implementable today on consumer hardware (RTX 4050, 6GB VRAM).

Porphyrins are π-conjugated materials with exceptional optical and electronic properties that have garnered considerable interest in the field of organic photovoltaics. Their well-established synthetic accessibility enables the construction of diverse molecular geometries through relatively few synthetic steps. This review highlights recent progress over the past decade in the rational design of porphyrin-based donor–acceptor architectures for third-generation organic photovoltaic devices. Emphasis is placed on the roles of molecular symmetry, π-extension, and donor–acceptor energy level alignment. Key structural features, including metal centers, side chains, and π-bridges, are examined in relation to their impact on device efficiency and operational stability. The influence of structural modifications on HOMO–LUMO energy levels, active layer morphology, and overall photovoltaic performance is critically assessed. • The properties of porphyrins make them good semiconductors in organic solar cells. • Molecular engineering in porphyrins, enable fine-tuning of energy levels. • Active layer morphology can be optimized by chemical functionalization. • Porphyrin-based semiconductors have evolved to multifunctional components in OSC. • The use of β-functionalized and n-type porphyrins for OSC is still challenging.

Distributed power (DP) generation through solar and wind energy holds immense potential in meeting future energy demands. Recent advancements in grid technology enable efficient utilization of local energy sources, minimizing transmission line losses and facilitating enhanced power distribution capabilities. This study focuses on examining the impacts of disturbances and faults within a four-area system linked to a Doubly Fed Induction Generator (DFIG) using discrete-time modeling. The primary control approach within the four-area system involving the DFIG incorporates various discrete-time Power System Stabilizers (PSSs). A comparative analysis of time-domain responses is conducted across various scenarios, including systems without controllers (no PSS), systems equipped with traditional PSSs such as Delta Omega, Multi-Band, Delta Pa and Fuzzy Logic (FL) based PSSs. A discrete-time mode PSS is specifically tailored for the silent-pole synchronous generator associated with the DFIG system. To assess the performance of the Fuzzy Logic (FL) controller in ensuring power system stability, a test DFIG model is employed. Time-domain simulations of the test DFIG model are carried out using the MATLAB/Simulink (R2015a) environment.

While Vehicular Ad-hoc Networks (VANETs) can potentially improve driver safety and traffic management efficiency (e.g. through timely sharing of traffic status among vehicles), security and privacy are two ongoing issues that need to be addressed. This paper proposes a novel blockchain-based authentication scheme leveraging conditional privacy within temporary platoons. The proposed model is an extended model based on Temporary Blockchain Platoons Formation. Our scheme utilizes blockchain technology to ensure data integrity and prevent unauthorized access, while enabling conditional privacy preservation where data sharing is limited based on pre-defined trust levels within the platoon. This significantly reduces reliance on Roadside Units (RSUs), leading to a cost-effective approach. Compared to traditional Public Key Infrastructure (PKI), our scheme offers improved security, enhanced privacy, and reduced infrastructure costs, promoting wider VANET implementation. The augmented model also incorporates an “Accident Prevention Mechanism,” a critical component absent in the preceding Platoon proposed model, thereby enhancing the overall safety and robustness of the system. Apart from the extra proposed features, rest features and mechanisms will be working as in already present Platoon Model.

The object of the study is the clinical decision-making process for selecting of hepatocellular carcinoma (HCC) treatment method based on the patient's medical data. The process of the HCC treatment method selection remains poorly formalized and is characterized by multi-criteria and the presence of numerous clinical situations, for each of which it is necessary to promptly identify the most accurate therapeutic solution. This study develops an intelligent medical decision support system for HCC treatment method selection based on knowledge applicable in clinical practice. It offers architectural and functional principles of the intelligent decision support system, classifying clinical situations by HCC treatment method, consisting of multiple possible combinations of 44 informative parameters. Based on the current values of these parameters, expert knowledge is transformed into production rules identifying HCC treatment methods. A heuristic procedure treatment selection is developed based on production rule analysis in accordance with current parameter values, reproducing the reasoning patterns of participants in a multidisciplinary council during their consensus decision-making process regarding HCC treatment appointment. A software implementation of a decision-making model for HCC treatment selection, implemented in C# using the Visual Study 2019 platform, enabled the integration of an intelligent system with web technologies. The intelligent medical decision support system automates the unique experience of professionals and helps physicians in a multidisciplinary consultation make prompt and informed decisions online regarding the appointment of personalized therapy. The system was piloted with expert physicians in several iterations until complete match between the consensus decision of the multidisciplinary council and the decision made by the developed system in accordance with clinical recommendations was achieved