Intel Labs

Agentic AI is quickly becoming the next evolution of AI. Learn more about the benefits and potential pitfalls in the blog coauthored by Intel and the National Institute of Standards and Technology (NIST). #AgenticAI #Innovation #Research

There’s a lot happening at this year’s International Solid-State-Circuits Conference [ISSCC] in San Francisco! Intel Labs researchers are presenting six papers and two university collaborations at the main conference, while also participating in forums and a special event panel. Plus, Navid Shahriari, Senior Vice President of Intel Foundry Technology Development, spoke about the AI Era Innovation Matrix, a matrix of technologies that are ushering in progress at every level, from chips to systems. Read the blog to learn more about these breakthrough technologies. https://intel.ly/3Qqk9yN #ISSCC #IntelXeon #Developer

While generative AI can create amazing works of art and even songs, using the internet as a dataset can cause copyright issues. Our Trusted Media Research Team is researching new ways to protect art and voices from potential unauthorized use. 🎨 Protecting art: This research tool prevents unauthorized use in AI-based generative image synthesis applications, securing ownership of image data. 🗣️ Protecting voices: This research tool uses a lightweight adversarial generation approach that learns protected voice versions of audio and confounds cloning techniques performed by other generative AI. Learn more about the research. https://intel.ly/40UhceD #AISecurity #GenAI #ResponsibleAI

Step into the lush and intricate Jungle Ruins, a sandbox environment built for integrating and stress-testing our research while providing a visually rich scene for creative exploration. Utilizing a combination of generative AI, procedural generation, and modular design, the Amazon rainforest-inspired Jungle Ruins took shape, all of which we offer to you under a Creative Commons license to help further enable the broader research and creative communities to experiment and build upon our work. Read the blog to learn more about how Jungle Ruins was created: https://intel.ly/40VLE8i #Developer #GenAI #DigitalArt #CreativeCommons

The era of better #AI is here — thanks to the power of an open ecosystem. The entire Intel product portfolio, along with our robust suite of developer resources, already supports hundreds of thousands of open-source AI models, including DeepSeek AI. #IntelXeon #IntelGaudi #IntelCoreUltra

Intel Scalable Vector Search (SVS) provides vector similarity search on billions of high-dimensional vectors with high accuracy, state-of-the-art speed, and a smaller memory footprint. Check out the blog to learn more about the advantages of our SVS performance library: https://intel.ly/3Cs2vaN #AI #VectorSearch #Developer

One of the most pressing challenges in the responsible use of AI is addressing bias in conversational agents. Training data may reflect societal prejudices around gender, race, ethnicity, and ideology, which can lead to model outputs reinforcing those stereotypes. To combat this, our researchers are exploring using large language models as judges to evaluate gender bias in models. Learn more about the method in this piece from Saurav Sahay and Shachi H Kumar. https://intel.ly/42BpNW5 #LargeLanguageModels #ArtificialIntelligence #ResponsibleAI

Evaluate the robustness of your GenAI language models with Intel Labs' Large Language Model Adversarial Robustness Toolkit, LLMart. The open-source toolkit offers scalable attacks, flexible configurations, and comprehensive support for Hugging Face models, and includes features like swap parallelization to enable efficient attack optimizations. Read this blog from Intel Labs research scientist Cory C. to learn more about LLMart and how you can protect your language models from adversarial attacks. https://intel.ly/40LdtRT #GenAI #AISecurity #OpenSource

We are committed to advancing AI technology responsibly. With rigorous multidisciplinary reviews, internal and external governance, and collaborative research projects, we are taking a holistic approach to growing AI initiatives that will drive positive change and better mitigate risks.   Learn more about our Responsible AI research: https://lnkd.in/ggWvbreg   #ArtificialIntelligence #Research #ResponsibleAI

ClimDetect is an open-source benchmark dataset for detecting climate change signals in daily weather patterns. Using vision transformers for image tasks, this novel approach shows great promise in analyzing spatial climate data, unlocking new opportunities for AI in environmental monitoring, and providing a physical grounding for AI-based insights. Read the blog from Sungduk Yu, Anahita Bhiwandiwalla, and Yaniv Gurwicz to learn more. https://intel.ly/4g6wZfZ #ClimateChange #OpenSource #VisionTransformers