Call for papers! The first workshop on Knowledge Augmented Methods for NLP (#NLProc) at #AAAI2023 is welcoming submissions🙌! Papers due on Nov. 4! Papers will be non-archival, so published papers (e.g.#EMMLP2022) can also present at our workshop! Details👉knowledge-nlp.github.io/aaai2023

𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗲 𝗥𝗮𝘁𝗵𝗲𝗿 𝗧𝗵𝗮𝗻 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗲 is now 𝗮𝗰𝗰𝗲𝗽𝘁𝗲𝗱 to #𝗜𝗖𝗟𝗥𝟮𝟬𝟮𝟯 🎉🎉 Without using DPR/Google, it achieved SoTA on multiple open-domain QA and knowledge-intensive benchmarks! Work done @ms_knowledgenlp! Code and paper: arxiv.org/abs/2209.10063

Combing Retrieval AND Generation (in step1) can further improve the model performance, as shown in Figure 3. The choice of retrieval or generation is interesting, and their complementarity is worth exploring. Using retriever or generator only where it helps.

🎉 Accepted papers (6 oral + 20 poster) are now online! Check them at knowledge-nlp.github.io/aaai2023/publi…

Thank you for the interest in our team! Sorry for not being able to reply everyone. We will keep hiring interns quarter by quarter. Feel free to message us your personal page or CV by Twitter whenever you are available. Our researchers will talk with you if matched : )

GPTEVAL: NLG Evaluation using GPT-4 with Better Human Alignment abs: arxiv.org/abs/2303.16634

The second workshop on Knowledge Augmented Methods for NLP at #KDD2023 is welcoming submissions📷! Papers due by May 23rd! Accepted paper will be non-archival! Details are available 📷 knowledge-nlp.github.io/kdd2023

THE BATCH selected four pieces of news after today's Andrew's newsletter, from Hinton leaving G to Wenhao Yu GenRead paper at ICLR. The work is open sourced. Wenhao is on the market! Grab him asap! deeplearning.ai/the-batch/issu… github.com/wyu97/GenRead

Do you know small models can be LLM's plugins? 🥳Introducing SuperICL, it's ICL but super dope! 🚀SuperICL combines small models with LLM like GPT-3.5 and can improve accuracy, multilinguality and interpretability! 📄Paper: arxiv.org/abs/2305.08848 🔧Code: aka.ms/SuperICL

Small Models are Valuable Plug-ins for Large Language Models abs: arxiv.org/abs/2305.08848 paper page: huggingface.co/papers/2305.08…

Any-to-Any Generation via Composable Diffusion present Composable Diffusion (CoDi), a novel generative model capable of generating any combination of output modalities, such as language, image, video, or audio, from any combination of input modalities paper page: