About
Articles by Prashant
Contributions
Activity
-
Day 18 of #30DaysOfFLCode Started reading the book: "Programming Differential Privacy" by Joseph P. Near and Chiké Abuah. It's a really good…
Day 18 of #30DaysOfFLCode Started reading the book: "Programming Differential Privacy" by Joseph P. Near and Chiké Abuah. It's a really good…
Liked by Prashant Kulkarni
-
Ilya Sutskever just confirmed at NeurIPS that pre-training scaling laws for LLMs have ended 👇 The compute is scaling but data isn’t (new or…
Ilya Sutskever just confirmed at NeurIPS that pre-training scaling laws for LLMs have ended 👇 The compute is scaling but data isn’t (new or…
Liked by Prashant Kulkarni
-
Every science practitioner will give due recognition to the great achevements of ancient science. But none will argue, that the old science is…
Every science practitioner will give due recognition to the great achevements of ancient science. But none will argue, that the old science is…
Liked by Prashant Kulkarni
Experience & Education
Licenses & Certifications
-
AI for Science on Supercomputers: Beginner
Argonne National Laboratory
Publications
-
Automatically Detecting Expensive Prompts and Configuring Firewall Rules to Mitigate Denial of Service Attacks on Large Language Models
Technical Disclosure Commons
Denial of service attacks on generative artificial intelligence systems, e.g., large language models (LLMs), can include sending LLMs requests that include expensive prompts designed to consume computing resources and degrade model performance. This disclosure describes techniques to automatically detect such prompts and then configure firewall rules that prevent such prompts in subsequent requests from reaching the LLM. Per the techniques, prompts provided to an LLM are matched against input…
Denial of service attacks on generative artificial intelligence systems, e.g., large language models (LLMs), can include sending LLMs requests that include expensive prompts designed to consume computing resources and degrade model performance. This disclosure describes techniques to automatically detect such prompts and then configure firewall rules that prevent such prompts in subsequent requests from reaching the LLM. Per the techniques, prompts provided to an LLM are matched against input and output token size as well as resource utilization to identify prompts that deviate significantly from a baseline. Expensive prompts are identified, and semantically similar prompts are automatically generated using the same LLM or another model. A subset of the generated prompts that are semantics similar to expensive prompts are identified by comparing respective vector embeddings. The subset of prompts and the received expensive prompts are provided to a pre-trained LLM that generates firewall rules, e.g., web application firewall (WAF) rules. Incoming requests from applications are evaluated based on the rules, and expensive prompts are blocked from reaching the LLM or are rate-limited.
-
Method to isolate tenancies for Large Language Modules (LLMs) Applications
Technical Disclosure Commons
Large language models (LLMs) and other types of generative artificial intelligence can be used in a wide variety of business applications. However, there is a possibility of data leakage from LLM responses when an LLM is used in shared multi-tenant environments where each tenant has respective private datasets. Deploying individual adapter layers for each tenant can provide data isolation. However, such implementations can be complex and costly. This disclosure describes techniques to create…
Large language models (LLMs) and other types of generative artificial intelligence can be used in a wide variety of business applications. However, there is a possibility of data leakage from LLM responses when an LLM is used in shared multi-tenant environments where each tenant has respective private datasets. Deploying individual adapter layers for each tenant can provide data isolation. However, such implementations can be complex and costly. This disclosure describes techniques to create and maintain a single model that can serve multiple tenants, with security controls for multi-tenancy services to isolate customer data efficiently. Data for different tenants is signed with their respective tenant-specific keys and is then appended with the tenant-specific signature prior to training/tuning a model or use by the model at inference time. When a business application of a particular tenant requests a response from the LLM, the response is generated using the adapter layer. The response includes data citations that are verified prior to the response being provided to the business application. The verification is based on the tenant-specific signature in the citation to ensure that only data that belongs to the particular tenant that requested the response is included.
-
Official Google Cloud Professional Security Certification guide
Packt
Google Cloud security offers powerful controls to assist organizations in establishing secure and compliant cloud environments. With this book, you'll gain in-depth knowledge of the Professional Cloud Security Engineer certification exam objectives, including Google Cloud security best practices, identity and access management (IAM), network security, data security, and security operations.
Other authorsSee publication -
A System and Method to Optimize LLM's Prompt Security Evaluation
Technical Disclosure Commons
Courses
-
Advanced Computer Vision with Deep Learning
ADSP 32023
-
Advanced Machine Learning & Artificial Intelligence (Tranformers)
ADSP 32017
-
Big Data Platforms
ADSP 31013
-
Data Mining Principles
ADSP 31008
-
Linear Algebra
ADSP 37016
-
Linear and Non-Linear Models
ADSP 31010
-
Machine Learning & Predictive Analytics
ADSP 31009
-
R for Data Science
ADSP 37020
-
Statistical Analysis
ADSP 31007
Languages
-
Hindi
-
-
Marathi
-
Recommendations received
7 people have recommended Prashant
Join now to viewMore activity by Prashant
-
♟️ Gukesh D: The Youngest World Chess Champion Ever! 🏆 In a moment that will be remembered in the annals of chess history, 18-year-old Gukesh…
♟️ Gukesh D: The Youngest World Chess Champion Ever! 🏆 In a moment that will be remembered in the annals of chess history, 18-year-old Gukesh…
Liked by Prashant Kulkarni
-
We are pleased to welcome Northwestern University as the 9th American institution to invest in the Giant Magellan Telescope. With institutions from…
We are pleased to welcome Northwestern University as the 9th American institution to invest in the Giant Magellan Telescope. With institutions from…
Liked by Prashant Kulkarni
-
🚀 Day 24: #30DaysOfFLCode Challenge by OpenMined. As, the #30DaysOfFLCode event is coming to end (Although, I have plan to continue ), I am…
🚀 Day 24: #30DaysOfFLCode Challenge by OpenMined. As, the #30DaysOfFLCode event is coming to end (Although, I have plan to continue ), I am…
Liked by Prashant Kulkarni
-
The Astrophysics and Observational Cosmology group at Newcastle University are delighted for Adam Ingram who will be hosting an ERC Consolidator…
The Astrophysics and Observational Cosmology group at Newcastle University are delighted for Adam Ingram who will be hosting an ERC Consolidator…
Liked by Prashant Kulkarni
-
Excited to attend #NeurIPS in Vancouver this year where we are presenting two works: 1. Scaling transformer neural networks for skillful and…
Excited to attend #NeurIPS in Vancouver this year where we are presenting two works: 1. Scaling transformer neural networks for skillful and…
Liked by Prashant Kulkarni
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Prashant Kulkarni in United States
-
Prashant K.
-
Prashant Kulkarni
Senior Engineering Manager - Business Applications at DoorDash ex Flexport | Atlassian | LinkedIn
-
Prashant Kulkarni
-
Prashant Kulkarni
-
Prashant Kulkarni
47 others named Prashant Kulkarni in United States are on LinkedIn
See others named Prashant Kulkarni