🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

What a way to celebrate one year of incredible Gemini progress -- #1🥇across the board on overall ranking, as well as on hard prompts, coding, math, instruction following, and more, including with style control on. Thanks to the hard work of everyone in the Gemini team and

Today we are rolling out an experimental Gemini Embedding model for developers with: – SOTA performance on MTEB (Multilingual) - Input context length of (3K --> 8K) tokens – Output 3K dimensions – Support for over (50 --> 100) languages More details in 🧵

The new model, gemini-embedding-exp-03-07, is our most capable text embedding model yet, surpassing our previous embedding model – text-embedding-004. It’s top ranked on the Massive Text Embedding Benchmark (MTEB) Multilingual leaderboard!

You can access the new experimental Gemini Embeddings model through the Gemini API right now, we plan to follow up with a production ready version in the months to come: developers.googleblog.com/en/gemini-embe…

Gemini Embedding Generalizable Embeddings from Gemini

🎉 Gemini Embedding is LIVE! 🎉 Try our state-of-the-art text embedding model for FREE on Vertex AI (text-embedding-large-exp-03-07; 120 QPM) & AI Studio (gemini-embedding-exp-03-07)! ➡️ APIs: bit.ly/gem-embed-vert…, bit.ly/gem-embed-aist… ➡️ Report: bit.ly/gem-embed-paper

Existing IR/RAG benchmarks are unrealistic: they’re often derived from easily retrievable topics, rather than grounded in solving real user problems. 🧵Introducing 𝐅𝐫𝐞𝐬𝐡𝐒𝐭𝐚𝐜𝐤, a challenging RAG benchmark on niche, recent topics. Work done during intern Databricks 🧱

Google folks continues to do awesome late interaction work. Compared to vanilla ColBERT, a version of this new “CRISP achieves an 11x reduction in the number of vectors—with only a 3.6% quality loss”.

CRISP: Clustering Multi-Vector Representations for Denoising and Pruning arxiv.org/abs/2505.11471 Multi-vector embeddings are better than single-vector on search/retrieval tasks but also have prohibitively more costly representations (e.g., ColBERT’s one embedding per token).

Did you know that fine-tuning retrievers & re-rankers on large but unclean training datasets can harm their performance? 😡 In our new preprint, we re-examine popular IR training data quality by pruning datasets and identifying and relabeling 𝐟𝐚𝐥𝐬𝐞-𝐧𝐞𝐠𝐚𝐭𝐢𝐯𝐞𝐬! 🏷️

Excited to present FastTD3: a simple, fast, and capable off-policy RL algorithm for humanoid control -- with an open-source code to run your own humanoid RL experiments in no time! Thread below 🧵

Dear MAGA friends, I have been worrying about STEM in the US a lot, because right now the Senate is writing new laws that cut 75% of the STEM budget in the US. Sorry for the long post, but the issue is really important, and I want to share what I know about it. The entire

R3-RAG: Learning Step-by-Step Reasoning and Retrieval for LLMs via Reinforcement Learning Introduces RL to teach LLMs adaptive search intensity scaling, i.e. dynamically adjusting retrieval depth based on problem complexity. 📝arxiv.org/abs/2505.23794 👨🏽‍💻github.com/Yuan-Li-FNLP/R…

🚨 Context matters for effective retrieval—but most embedding models cannot leverage crucial information outside of the passage they embed. Our new paper "Context Is Gold to Find the Gold Passage" explores how context-aware embeddings can be trained to boost performance! 🧵(1/N)

🚀 ColQwen2 just dropped in Transformers! 🤗 Say goodbye to brittle OCR pipelines — now you can retrieve documents directly in the visual space with just a few lines of code. Perfect for your visual RAG workflows. Smarter, simpler, faster. Let's dive in! 👇 (1/N 🧵)

I'm thrilled to announce the release of FastPlaid ! 🚀🚀 FastPlaid is a high-performance engine for multi-vector search, built from the ground up in Rust (with the help of Torch C++)⚡️ You can view FastPlaid as the counterpart of Faiss for multi vectors.

Multimodal RAG: Just use ColPali/DSE then pass your screenshots to the LLM This is the dream, but how well do LLMs read text contained in images? We wanted to know, so we tried a simple thing: do results change on evals when using screenshots rather than text as input? Yes.

Excited to share our new work, CLaMR! 🚀 We tackle multimodal content retrieval by jointly considering video, speech, OCR, and metadata. CLaMR learns to dynamically pick the right modality for your query, boosting retrieval by 25 nDCG@10 over single modality retrieval! 🧐

Video-ColBERT: Contextualized Late Interaction for Text-to-Video Retrieval Introduces a bi-encoder approach that performs fine-grained token-wise interaction at both spatial and temporal levels using modified MaxSim operations and dual sigmoid loss. 📝arxiv.org/abs/2503.19009

Neural embedding models have become a cornerstone of modern information retrieval. Today we introduce MUVERA, a state-of-the-art retrieval algorithm that reduces complex multi-vector retrieval back to single-vector maximum inner product search. More →goo.gle/4k8YRlN