📈 Why it matters: This opens the door to smarter educational tools, better automated diagram explanations, and richer multimodal AI assistants that can reason across text and geometry. #AI #MachineLearning #DataScience #ComputerVision #GeometricAI

💡 For context, this means that the model doesn’t just see a triangle. It can determine and state: “an equilateral triangle with all sides equal.”

🔹 Used Reinforcement Learning with Verifiable Rewards (RLVR) to align captions with images. 🔹 Training on this data boosts multimodal models’ cross-modal reasoning, with strong results on benchmarks like MathVista and MathVerse.

🚀 From Geometry to Words New research introduces Geo-Image-Textualization (GGICS) — a method for generating high-quality image–caption pairs in geometric problems. 🔹 Built the GeoReasoning-10K dataset (10,000 pairs).

👉 Neural models are bringing acoustics out of the lab and into ML workflows. 💭 What are your thoughts? #AI #MachineLearning #AcousticModeling #RIR #SpeechEnhancement #VirtualAcoustics

𝐖𝐡𝐲 𝐢𝐭 𝐦𝐚𝐭𝐭𝐞𝐫𝐬: ⚡ Much faster than full simulations 🎯 Controllable & realistic ✅ Validated with acoustic metrics like T60 & DRR Applications: cleaner speech models, de-reverberation, immersive AR/VR.

Until now, simulating those echoes required heavy physics engines or big measured datasets. However, new research (NACF, AV-RIR, MiNAF) indicates that we can now generate RIRs from simple contexts, such as room size, wall materials, and the speaker/listener's position.

Neural Acoustic Models Meet Real-World RIRs 🔊 Imagine clapping your hands in a cathedral vs. in your kitchen. The echoes are completely different....that’s the room impulse response (RIR).

Ça fait longtemps ici.

Hey plant/microbe science world!🌱 We are looking for an enthusiastic PhD candidate in the Green Life Sciences in Amsterdam, Netherlands. Are you interested in plant microbiomes and how they help alleviate stress? Apply here👇 https://t.co/tD13iUhusP @SILS_UvA @MiCRop_research

🚨 PhD position alert! 🚨 I'm hiring a fully funded PhD student to work on mechanistic interpretability at @UvA_Amsterdam. If you're interested in reverse engineering modern deep learning architectures, please apply:

postdoc position available in Lisa Chakrabarti's lab at @institutpasteur in Paris. @Team_Chakra topic: organs-on-ship to test mRNA vaccines see pic for details:

Hey! I am looking for several postdocs and engineers to work with me on a DARPA project to start ASAP. Looking for ML, Robotics, or CV skills. Here is a video about the project: https://t.co/yllyYE1SWf If that's interesting for you, consider applying:

I am hoping to recruit a Ph.D. student to join our Social Injustice and Social Change research group at @UCSBpsych in Fall 2025! You can find more information about the position, about our team, and about our research on our website.

We are hiring! I am looking for a postdoc for a project in mechanistic modeling who is not afraid to learn about spectroscopic measurements of photosynthesis. Please spread the word! More information here:

We have 15 PhD positions available as part of the DANIO-ReCODE MSCA DN. Great opportunity for anyone interested in developmental biology, regeneration, genomics, and computational sciences. For project descriptions and how to apply, check:

𝗨𝗽𝗱𝗮𝘁𝗲 𝗮𝗻𝗱 2𝗣𝗵𝗗 𝗽𝗼𝘀𝗶𝘁𝗶𝗼𝗻s 𝗶𝗻 𝗨𝗿𝗯𝗮𝗻 𝗰𝗹𝗶𝗺𝗮𝘁𝗲 Starting January 2025, I’ll join the Department of Geography at Texas A&M University as an Assistant Professor. I am currently recruiting two motivated graduate students to join my lab - see below

My team at Google is looking for a full-time research scientist to work on VLM and Multimodal. Interested candidates can apply directly at https://t.co/50QmOJXIeW.

I’m advertising several PhD projects in palaeoclimate, stratigraphy & palaebiology of past greenhouse worlds! Funding available via our new @NERCscience @BBSRC doctoral programme: https://t.co/M5CBxrQQ6m Projects described here:

🧐I am recruiting Postdoc and PhD students @mcgillu @Mila_Quebec ! PhD on Fairness and Privacy: Apply through @Mila_Quebec: https://t.co/ccTPHuaT0m Postdoc on Machine Unlearning in Text-to-Image Models: Send me your CV! EQUAL Lab: https://t.co/j6cI24FBTk ⏰Deadline: Dec 1st