Weekly AI reading and watch list - Nov 1st, 2024
Fascinating AI articles, papers, and books I discovered this week
Every Friday, I aim to share a list of AI papers, blog articles, books, and videos that I find worth reading or watching. While many will be recent, I'll also include older but equally, if not more, significant works.
LLM Agents Hackathon
The LLM Agents Hackathon, hosted by Berkeley RDI (https://rdi.berkeley.edu/) in collaboration with the LLM Agents MOOC (https://llmagents-learning.org/f24), is a unique opportunity to participate together students, researchers, and AI practitioners to innovate and expand the frontier of LLM agent technology. Whether you're diving into building with LLMs for the first time or an experienced developer, this is the place to share, grow, and be inspired by cutting-edge work in AI agents.
🗓️ Open to the public, this hackathon offers both virtual participation and an in-person event at UC Berkeley! Join to contribute to the future of AI and network with like-minded individuals in the AI space. You can win more than $200k in prizes and resources!
Ready to innovate? You can apply here: https://rdi.berkeley.edu/llm-agents-hackathon/.
Berkely LLM Agents MOOC
The public lecture series on LLM agents by Berkeley RDI, related to the hackathon mentioned above, has already begun but remains an excellent resource for gaining an up-to-date and in-depth expertise in LLM agents.
COMPL-AI is an open-source compliance-centered evaluation framework for Generative AI models
COMPL-AI provides a first Technical Interpretation of the EU AI Act. It consists of a detailed, technical breakdown of the high-level regulatory requirements outlined in the EU AI Act. Requirements are categorized under six guiding principles of the Act
Transparency
Social & environmental well-being
Human agency and oversight
Diversity, non-discrimination and fairness
Privacy and data governance
Technical robustness and safety
and labeled based on risk levels:
General Purpose AI (GPAI),
GPAI with Systemic Risk (GPAI+SR), and
High-Risk (HR).
Website
Paper
Evaluate your own LLM and compare it on the EU AI Act Compliance Leaderboard
COMPL-AI also offers an open-source benchmarking suite, which allows to evaluate your own LLM and compare it against 27 state-of-the-art LLM benchmarks. Both the benchmark suite and technical interpretations are open for community input and contributions.