Blazars exhibit complex broadband emission driven by various physical processes. Self-consistently fitting their SEDs with numerical models is computationally intensive, requiring significant resources and time. In thisn project, using Convolutional Neural Networks (CNNs), we aim to train these networks on both leptonic and hadronic radiative models, which represent the physical mechanisms underlying blazar emissions. These models account for all relevant cooling processes of electrons/positrons, protons, neutrons, and secondary particles generated through photo-pion and photo-pair interactions, providing a detailed description of particle interactions within the jet environment and the resulting radiation. Once trained, the CNNs efficiently reproduce the radiative signatures of these models, significantly reducing computational time  without compromising accuracy. By converting complex radiative frameworks into efficient neural network tools, this approach offers a powerful method for studying the physical conditions within blazar jets and exploring the complex parameter space that governs these high-energy phenomena. Furthermore, the methodology is versatile and can be adapted to other astrophysical sources, provided a representative set of spectra is available for training. This marks a significant step forward in integrating machine learning into astrophysical modeling.

Key features of AstroGenesis include: 

• Literature Intelligence: Analyzes and retrieves relevant scientific publications to provide context-aware insights grounded in peer-reviewed research.  

• Astrophysical Data Access: Enables direct access to multiwavelength and multimessenger observations through connections to astrophysical catalogs and databases such as the Markarian Multiwavelength Data Center (MMDC).  

• AI-Driven Modeling: Supports the modeling of astrophysical sources using machine-learning representations of physical simulations, enabling rapid exploration of parameter space.

• Integrated Research Workflow: Combines literature analysis, data retrieval, and modeling within a single interactive environment to streamline the research process.

• Education and Exploration: Serves as a tool for students and researchers to explore astrophysical concepts, datasets, and models in an intuitive way. 

AstroGenesis represents a step toward AI-assisted scientific discovery, where intelligent systems support researchers in navigating complex data, interpreting results, and developing new insights into the high-energy universe.

Its key functionalities include:  

• Data Access: Automated retrieval of data from multiple sources, covering all energy bands from radio to γ-rays, ensuring comprehensive datasets for blazar studies.

• Interactive Visualization: Tools for plotting multi-band SEDs with various filters, allowing detailed temporal and spectral analyses.

• Theoretical Modeling: Advanced modeling capabilities utilizing CNNs trained on self-consistent leptonic and lepto-hadronic models. This feature significantly reduces computational time and enables precise parameter estimation and interpretation of observed data.

The integration of comprehensive datasets, theoretical modeling tools, and advanced AI capabilities positions MMDC as an indispensable resource for studying the complex physical processes driving blazar emissions.