I am excited to share our #CVPR2024 DEF-MIA paper on the Interactive Medical Image Learning Framework (IMIL) 🩻!

We showed that IMIL could effectively use clinician feedback on just 4% of the dataset to yield a 4.2% improvement in accuracy! 🚀

🧵 below
arxiv.org/abs/2404.10965

Dolphin-2.9-8x22b is in the oven.
fft, deepspeed zero3 param offload, 8k sequence, half the layers are targeted.
This is a significantly improved, filtered dataset. Function calling, agentic, math, dolphin and dolphin-coder.

Boost you job search profile in data analytics with these project dataset:

How do I make myself a dataset without organs?

Build a strong data science and analytics portfolio from these best end-to-end projects:

- Dataset are provided.
- Full explanations are given.

‼️ Major problems identified in the data adjustments applied to a widely used global temperature dataset ‼️

A study from 2022 showed that the homogenization algorithm NOAA uses to adjust temperature data is inconsistent and does not correspond to reality.

1 dataset, two dashboards. A or B?

#30DayChartChallenge | day 17- networks

African and Eastern Refugee Migration from 2010 to 2022 using population dataset from the refugee R package

#rstats #dataviz #Datavisualization #migrants #Africa #MiddleEast

AINN Layer2 Dataset Phase 2 is now live! #PIIN #ANVM #AINN

ANVM | AINN Layer2 went live on the mainnet on March 31st, initiating a staking campaign in collaboration with #PIIN community. The staking activity will run for 2-3 months, with participants receiving future ANVM tokens from…

2️⃣ It's time for Phase Two! 2️⃣

🧵 1/5

📢 Announcing the 2nd phase of AINN Layer2's Dataset Staking Event!

🟣 Enjoy 1️⃣ Month of 0⃣ Bridge Fee for BTC/BRC20 inscription assets starting today! Don't miss out!

5 overlooked ChatGPT prompts that R data scientists can use to save 10+ hours per week:

1. Generating SQL Queries:

Prompt: 'Given a dataset description with columns such as customer_id, purchase_date, and amount_spent, can you help me construct a SQL query to find the total

#echofirst 👀3D TEE Basic Dataset👉just 1️⃣2️⃣3️⃣ 2D reference view needed
🔻switch off the respirator
🔻use multibeat (4-6)
🔻aim > 15 FPS
#CardioTwitter ⤵️

There’s simply no way to guarantee that what gets spat out of an AI model isn’t copyright infringement. These models are trained on millions of images and there’s no way to verify if the image you’ve just generated hasn’t taken way too much from an existing work in the dataset.

The result of processing a dataset generated by artificial intelligence and additionally interpreted in 3d space.

#sculpture #light #design #technology #sidefx #houdini

Just found out that pond5 automatically opts in people's work for sale to generative AI companies without telling them.

They send out an email to say participation in the Dataset Program is optional, but it only happens *after* a contributor's data has been sold. 1/

ECCV 2024 has officially accepted our proposal to host the 1st Dataset Distillation Challenge. This event will be co-hosted by 19 renowned universities and research institutions from 7 countries, including MIT, Stanford, and Google.

Beaten by a Commodore 64.

this post was made by a Quantum AI (which uses X as a training dataset)
tomshardware.com/tech-industry/…