#NatureLM
LightNews LightNews
zendarva.bsky.social
One of the better examples I've seen recently was I arxiv.org/html/2411.07...

This model has generalized bird calls from one species to another amongst other examples.
NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics
arxiv.org
October 10, 2026 at 4:26 PM
earthspecies.bsky.social
🌟 New Community Spotlight 🌟

Researchers used model merging to improve NatureLM-audio’s instruction-following while retaining its expertise in animal sounds.

This resulted in a 200%+ relative improvement in F1 score on unseen taxonomic families!

🔗 Learn more: earthspecies.org/2026/09/30/w...
What if a Model Knows More Than it Can Say?
Discover how researchers used model merging to unlock capabilities already present in NatureLM-audio and improve its ability to generalize to species it had never encountered during training.
earthspecies.org
October 9, 2026 at 1:34 PM
aichina.news
MIT-licensed MoE on Modelers.cn instead of Hugging Face is the real signal. But no evals, no usage notes, and a creation date still in the future — hard to see this displacing Qwen on Ascend until someone posts real numbers.
Microsoft's NatureLM-8x7B Appears on Modelers.cn: An MIT-Licensed MoE with Ascend Ambitions
aichina.news
September 18, 2026 at 6:46 PM
aichina.news
MIT licence on an 8x7B science MoE is unusually permissive — rare in this space, and Ascend-targeted too. But the card ships with no benchmarks, no training data, no hardware notes. Hard to see this displacing Qwen on a lab pipeline until someone publishes scores.
NatureLM-8x7B-Inst: A Science-Focused Mixtral MoE Lands on Modelers.cn
aichina.news
September 18, 2026 at 6:11 PM
aldesinaga.bsky.social
#MachineLearning
#ArtificialIntelligence #ConservationTech #NatureLM #TalkingToAnimals #WhaleTalk #NatureTech #DecodingNature
#WildlifeScience #Solomon #Salomo #Sulaiman #Jedidiah #Yedidiah
May 18, 2026 at 5:40 AM
aldesinaga.bsky.social
2. Leading Projects & Breakthroughs
Several specialized research organizations are leading the charge across different species:
The Earth Species Project (ESP)
An international nonprofit at the absolute forefront of this movement. They are building NatureLM-audio, the first
May 18, 2026 at 5:40 AM
kriptabiz.bsky.social
Новая эра научных открытий: как NatureLM помогает расшифровать язык природы

https://kripta.biz/posts/5061554A-0185-45A9-B3D3-F744C98FF8C8
September 19, 2025 at 5:32 AM
lodweb.eurosky.social
NatureLM: Deciphering the Language of Nature for Scientific Discovery arxiv.org/html/2502.07...
NatureLM: Deciphering the Language of Nature for Scientific Discovery
arxiv.org
February 13, 2026 at 1:54 PM
animal-prattle.bsky.social
Note they also apply NatureLM-Audio to their benchmark BEANS, IIUC

ieeexplore.ieee.org/document/100...

github.com/earthspecies...

bsky.app/profile/arxi...

#prattle 💬
November 17, 2024 at 2:05 AM
animal-prattle.bsky.social
icymi, Earth Species Project released a large language model (a la ChatGPT) you can prompt with text and audio:
arxiv.org/abs/2411.07186

we shared on the bird app, don't think any ESP core ppl are here yet (maybe @loganjames.bsky.social knows?)

Will big tech labs follow suit?
#prattle 💬
November 17, 2024 at 2:02 AM
grethchen.de
Deng gehört zu denen ihrn Naturelm derzu, ja odr?! 🤭☺️
April 19, 2026 at 9:59 PM
theaitrack.bsky.social
🌍 NatureLM-audio: Decoding Animal Sounds with AI 🐦🎶 The first AI model decoding animal sounds, from bird calls to whale songs. Zero-shot learning, species classification, and even bioacoustic captions!✨

More: tinyurl.com/mspe8j6v

#AI #AIForNature #Bioacoustics #ConservationTech
March 24, 2025 at 12:47 PM
researchtrend.ai
[2025-07-01] 📚 Updates in #AuLLM

(1) <a href="https://researchtrend.ai/papers/2506.23670" class="hover:underline text-blue-600 dark:text-sky-400 no-card-link" target="_blank" rel="noopener" data-link="bsky">Efficient Interleaved Speech Modeling through Knowledge Distillation
(2) Efficient Interleaved Speech Modeling through Knowledge Distillation
(3) NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics

🔍 More at researchtrend.ai/communities/AuLLM
July 1, 2025 at 3:58 AM
researchtrend.ai
(34) Ambient AI Scribing Support: Comparing the Performance of Specialized AI Agentic Architecture to Leading Foundational Models
(35) Cancer-Answer: Empowering Cancer Care with Advanced Large Language Models
(36) NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics
November 22, 2024 at 12:22 PM
researchtrend.ai
(39) Ambient AI Scribing Support: Comparing the Performance of Specialized AI Agentic Architecture to Leading Foundational Models
(40) Cancer-Answer: Empowering Cancer Care with Advanced Large Language Models
(41) NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics
November 21, 2024 at 10:45 PM
andresgarciaramon.bsky.social
www.tiktok.com/@aidrugsyrob...
🦁🎵 NatureLM-audio: El primer modelo de IA que nos acerca a entender el lenguaje animal 🐘🔊 #ia #ai #naturelm #animals #animales #hablar #audio #entender #lenguajeanimal #artificialintelligence #intelig...
TikTok video by AI, Drugs & Robots 🧠
www.tiktok.com
December 25, 2024 at 10:21 PM
sondervorst.bsky.social
La force de l’IA, c’est d’analyser de grandes quantités de données qui échappaient jusqu’ici à notre analyse.

Et Google n’est pas seul dans la course : l’Earth Species Project a reçu récemment des fonds du cofondateur de LinkedIn, Reid Hoffman, pour développer son modèle NatureLM.
Earth Species Project
www.earthspecies.org
April 19, 2025 at 4:15 PM
nic221.bsky.social
Introducing NatureLM-audio: An Audio-Language Foundation Model for Bioacoustics - Earth Species Project https://www.earthspecies.org/blog/introducing-naturelm-audio-an-audio-language-foundation-model-for-bioacoustics #AI #animals
Text Shot: Trained on a newly compiled dataset combining large bioacoustic archives, human speech, and music, this model can solve a wide range of bioacoustics tasks zero-shot by simply being prompted with natural language queries and audio. From this input, it generates free-form text answers, greatly enhancing usability. For example, NatureLM-audio can classify or detect thousands of species across diverse taxa including birds, whales, and anurans – without the need to retrain the model for each new task and without machine learning and programming expertise. Evaluating on a novel benchmark, BEANS-Zero, we find NatureLM-audio can complete multiple tasks previously unexplored at a cross-species level, such as predicting life-stage and simple call-types of birds, and captioning bioacoustic audio. These new capabilities show promise to accelerate data processing for communication studies across a wide range of species.
November 18, 2024 at 4:10 AM
nic221.bsky.social
NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics https://earthspecies.github.io/naturelm-audio-demo/ #AI #bioacoustics #EarthSpeciesProject #animals
Text Shot: these capabilities have yet to be fully demonstrated in bioacoustics tasks, such as detecting animal vocalizations in large recordings, classifying rare and endangered species, and labeling context and behavior—tasks that are crucial for conservation, biodiversity monitoring, and the study of animal behavior. In this work, we present NatureLM-audio, the first audio-language foundation model specifically designed for bioacoustics. Our carefully curated training dataset comprises text-audio pairs spanning a diverse range of bioacoustics, speech, and music data, designed to address the challenges posed by limited annotated datasets in the field. We demonstrate successful transfer of learned representations from music and speech to bioacoustics, and our model shows promising generalization to unseen taxa and tasks.
November 13, 2024 at 3:34 PM
cryptoknightai.bsky.social
最近阿里ATH-Token Foundry联合人大高瓴AI学院开源科学基础模型LOGOS，打通各学科数据孤岛

这次独创统一离散Token编码体系，蛋白质、小分子、材料等异构结构共用一套“科学语法”，不靠3D几何网络就能还原空间结构

仅1B参数量便超越微软NatureLM，无需微调即可投入科研使用，配套448.7亿token多模态科研数据集完整开源，重塑跨领域AI科研的底层技术范式
June 18, 2026 at 1:42 PM
NatureLM-audio: Decoding Animal Sounds
Researchers have introduced NatureLM-audio, an audio-language foundation model tailored for bioacoustics. This model excels in zero-shot classification, identifying species it hasn't encountered before, and interpreting animal vocalizations.
buff.ly
January 13, 2025 at 4:30 PM
romanian-news.bsky.social
Cercetătorii dezvoltă un sistem AI care va permite oamenilor să „vorbească” cu animalele
Cercetătorii dezvoltă un sistem AI care va permite oamenilor să „vorbească” cu animalele
Earth Species Project, o organizație non-profit, a prezentat NatureLM, un model de inteligență artificială (AI) menit să decodifice comunicarea dintre animale, transmite Firstpost. Această tehnologie…
petscats.ro
January 2, 2025 at 8:00 PM
kokuma.bsky.social
IA bien: este modelo está entrenado con sonidos de animales y responde a preguntas sobre grabaciones de campo: ¿qué animal suena? ¿es adulto? ¿cuántos hay? ¿es canto o llamada?

De momento hay paper y ejemplos, pero los autores prometen compartir el código.

earthspecies.github.io/naturelm-aud...
NatureLM-audio: an Audio-Language Foundation Model for Bioacoustics
earthspecies.github.io
November 27, 2024 at 8:05 PM
earthspecies.bsky.social
In 2025, we open-sourced NatureLM-audio, launched a multi-year study of killer whale communication with @raincoast.org, expanded collaborations across species, and co-organized the first AI for Non-Human Animal Communication workshop at NeurIPS.

Learn more: earthspecies.org/annual-repor...
June 30, 2026 at 3:45 PM