Do we have anything newer than #WordNet?

讲个题外话，英语入门后推荐在桌面安装个离线版 en-en 词典，推荐两个开源免费的：wordbook 小巧方便，数据来源于 wordnet。另一个是 kiwix，严格来说它是一个维基数据库浏览器，安装之后用它下载 wiktionary（9GB，可用磁力链下载）就可以象翻维基一样有趣地翻单词了。另：不推荐双语（双向）词典

A new, sophisticated jailbreak technique known as Neural Carrier Articles embeds prohibited queries into benign carrier articles in order to effectively bypass model guardrails. Using only a lexical database like WordNet and composer LLM.

We are in Global WordNet Conference 2025! https://t.co/lmwMTu2Y7P

2024/07/29 ・掲載語彙を大量に追加しました。現在147582の語彙を掲載中です。これにより、WordNet(上位語や下位語 など、意味の繋がりの辞書)の語彙間のリンクが全て機能するようになりました。 ・「もっと覚える」のコラムを100ほど追加しました。 ※イメージ画像や語源の情報は適宜追加予定です。

Machine translators compete in translating the Princeton WordNet concepts to Ukrainian 😀 #unlpworkshop #lrec2024

#SoftwareDevelopment #nlp #nltk #ai #Python I've been playing around with Natural Language Processing in #Wordnet and have created a site that lets you "Enter a word to search all the Synsets in Wordnet containing all four Parts Of Speech (POS), along with their corresponding

소설 쓸 때 유용한 구독 사이트를 소개합니다. (이미 다 아실듯) 단어를 입력하면 유의어와 관련어를 출력해주는 사이트입니다. 한국어 워드넷 사전이에요. https://t.co/iSvcbQzfR9 저는 한글창, 국어사전창, 워드넷 사전창. 이렇게 세 개 켜놓고 작업을 시작해요🤩

Hierarchical Retrieval with Dual Encoders • DEs can solve HR if dim. scales with depth/log(# docs) • Challenge: lost-in-the-long-distance — far ancestors hard to retrieve • Solution: Pretrain → finetune on long-distance pairs • Boosts recall on WordNet from 19% → 76%

Using Chat GPT reminds me of being 12 yrs old again. When the coolest thing to do online was talk to your friends on MSN & talk to @UltraHal- a chat bot. Hal won the ‘07 Loebner Prize for "most human" chatterbot (as they used to be called) & utilized wordnet, an early NLP tool.

A nicely presented paper with an interesting message: Categorical concepts in LLMs are represented as simplices. Hierarchically related concepts live in orthogonal subspaces. The findings align with the structure in the WordNet hierarchy.

Petya Osenova is presenting our paper "Recent Developments in BulTreeBank-WordNet" at Global WordNet Conference 2023, San Sebastian, Spain.

An automated way of building the Ukrainian WordNet from Princeton WordNet and WikiData 💡 Find details in https://t.co/zKjNNaojPV #unlpworkshop #lrec2024

> WordNet 😑

🚀 Exciting research alert! 🚀 GSSR-GNN: A model for constructing geospatial service similarity relationships🌍🔗By leveraging service descriptions, tags, and #BERT & #WordNet, it boosts service replacement efficiency & cuts computational costs🛠️💡 #AI #geospatial #Innovation

ウホようございます🦍 ボカデュオ最終日 七夕ソングで参加してます Last Starlight 歌い手版 https://t.co/QeMIwuybHp #VocaDuo2024 ボカロ版 https://t.co/vwyZ7FAvnG #VocaDuo2024vv ーーーー 過去に作った七夕ソングもぜひ！ wordnet https://t.co/3MBKcpcL4e 7月7日 https://t.co/6Zwo9P0rp2

Some LLM trainers have been studying ImageNet (WordNet) pretty intensively, some others not so much. The OpenAI one was expected, since it was already explicit part of CLIP's data construction, but I didn't expect it to get every single ID exactly right.

The Global WordNet Association Meeting has officially started!

there is an "etymological wordnet" and the dataset is free. however looking through it, it's not clear that the precise relationships we need are represented in it; example screenshot of a word that came to english through french (should be an edge here to an "fra:" node)