My name is Neetu Gangwani. Today I'm going to discuss about enhancing IoT security and privacy in cloud environments using blockchain federated learning for cross cloud trust and compliances. So let's get started. In today's world, when we talk about the technologies and software industries, everything is dependent on data, whether you deal with investment banking, healthcare, or inventory management systems, or ERPs, or anything else. So when we talk about the data, by default, something comes into our mind, that's the risk. Whether our data is secure or not, is it okay to share the data? Or what are the challenges when we talk about the privacy and security, especially when we are dealing with a distributed AI system. So to how we can overcome it, how we can enhance the security. So let's first discuss the problem challenges in privacy and security in distributed AI. First of all, data sensitivity, distributed data sets often contain sensitive personal and organizational data that cannot be shared or centralized. We can't share our personal data with anywhere like we don't want to share it and other security threats. AI systems are vulnerable to adversarial attacks like model poisoning and Sybil attacks compromising system integrity. Trust defects, collaboration across organization or cloud providers, lack of transparency leading to mistrust in the AI outputs, regulatory pressures, stricter data protections law, such as JDPR, HIPAA, require systems to ensure privacy and accountability. So what are the solutions? So there are other solutions. We can use a different methodology. To combine all these things to make the system more reliable, secure and private. So first one is a blockchain, a decentralized leisure, provide transparency, trust and immutable for AI model updates. Another one is federated learning and enable collaborative model training without sharing raw data, maintaining data privacy. Cloud AI combines the scalability and efficiency of cloud platform with privacy and preserving mechanism. let's just see each and every terminology in detail. So first one is a blockchain. What exactly meaning of the blockchain? So blockchain is basically a decentralized distributed ledger technology that ensure transparency, immutability, security of the transition. So it uses the cryptographic technologies and to create a chain of blocks Each contains a set of transaction key feature. Key features include decentralization with no single entity has control over the entire network. Transparency as all transactions are visible to the network participants and immutable, making it extremely difficult to alert recorded data. So in blockchain, what exactly happens? For example, user is sending one sentence and then over the network it will block, it will create the, it will break the sentence into the different blocks and each block will be combined after receiving by the participant. And these blocks, have the private keys and the, and they are getting transferred by using the different networks. Next, what is the federated learning? Federated learning is a machine learning paradigm that was built by Google in 2017 that enables training models to be distributed dataset without centralizing the data. In federated learning, multiple parties collaborate to train AI shared models while keeping their data locally, addressing privacy concerns, regulatory requirements, etc. The process typically involves central server coordinating the learning process, where local models are trained on individual datasets and only model updates are shared with the server. The server then aggregates these updates to improve the global model, which is redistributed to the party server. participants for the next round of training. Next, Cloud AI. Cloud AI refers to artificial intelligence service and infrastructure provided through cloud computing platform. It encompasses A wide range of offering from pre trained models and API to fully managed machine learning platforms. The current landscape is dominated by major cloud providers offering scalable AI solution, enabling our organization to leverage advanced AI capabilities without significant upfront investment in the hardware and the expertise. So what are the key components? Blockchain, when we combine these two technology into the framework, we are getting the multiple and advantages like decentralized model updates. Each update is treated as a transaction enhancing transparency traceability. So when we record each and every transaction, who has performed it, how performed, and these transactions should be immutable so that user don't have the rights to remove it, updated or do any modification. In the logs, so that's the beauty of the system moving to the next one smart contracts Automates validations and consequences for model updates, ensuring integrity. If something is updated at the one place, the system has to make sure all the copies are updated with the correct data and none of the copies have the older data. Incentive mechanisms reward high quality data contribution and prevent system gaming. Access control cryptographic identity management ensures only authorized. Participants contribute. For example, when you are dealing with the mobile at that time, either you use your mobile passcode or your fingerprints or your face ID. So these type of, Authentications are known as the cryptographic identity, then audit rail, immutable record of all the interactions for accountability and the compliances, benefits, advantage of integrated approach, enhanced Privacy keeps data localized while securely sharing model updates. So whatever the base model you have, that will be kept on your on prem servers or the centralized servers without having the connection with the internet. Improved trust, immutable records, build confidence in collaborative AI. cross cloud interoperability, enables seamless use of diverse cloud AI service, resistance to attack, combines block check consequences and federated learning, decentralization for robust security, regulatory compliances, providing transparency, audit drills to simplify JDPR and the HIPAA compliance, potential applications, so there are many. Many applications where we can enhance the privacy and security using blockchain enabled federated learning. For example, healthcare collaborates securely on patient data for drug discovery and personalized treatment. Finance trained fraud detection models across institutions without exposing sensitive transaction data. Smart cities optimize urban planning and resource management with privacy preserving AI models. Supply chain improves forecasting and inventory management without compromising business data. Age computing enabling AI. Enable iot devices to collaboratively train AI model keeping data secure and localized Comparative advantages why this approach is stand out. So first of all is a privacy preservation Techniques like differential privacy and homo Homomorphic encryption prevents sensitive data exposure. Select scalability layer 2 blockchain solution handle high transaction volume and distributed updates. Regulatory alignment designed to comply with JDPR, HIPAA and other global regulations. Collaboration innovation incentive structure encourage multiple party participation while preventing manipulations. Challenges and future direction. Addressing challenges and unlocking opportunities. Scalability. Research efficient algorithms like sharding to handle high volume systems with a great response time. Advanced privacy, implementing secure computational techniques to prevent, to protect even aggregated data. Incentive device, balancing rewards for data contribution, introducing, without introducing bias. Performance optimization, reducing computational overhead for blockchain integration in the federated learning. Governance, exploring DAOs for decentralized model and version control. Let's see a case study. For example, healthcare use care secure federated learning problem sharing data across sharing patient data across hospital is limited due to privacy concern and regulatory buyers solution federated learning enable hospital to train AI models to local data and blockchain secures model to ensure the transparency outcome. Improved diagnostic accelerated drug discovery and compliances with the GDPR and the HIPAA conclusion In the integration of blockchain technology with the federated learning in cloud AI environment represents a trans made transformative step forward in addressing critical challenges of privacy, security, and trust in distributed AI systems while leveraging the dis decentralized and immutable feature of blockchain combined with the privacy preserving capability of federated learning. This framework enables secure and collaborative AI model development without compromising sensitive data. Organization can now train model across distributed data set while ensuring compliances with regulations like GDPR, HIPAA fostering innovation in highly regulated industries such as healthcare, finance, and smart cities. Thank you.