Analytics DeepLearning.AI - The Batch | DeepLearning.AI (Page 47)

Mochi-style illustrated characters with diverse facial expressions used for AI emotion recognition visualizations.

Machine Learning Research

Interactive Voice-to-Voice With Vision: MoshiVis adds image understanding to voice-first conversations

Researchers updated the highly responsive Moshi voice-to-voice model to discuss visual input.

Scientists examining laptop through microscope in high-tech lab, researching data with advanced technology tools.

Data Points

Adapting R1-like techniques to video reasoning: Anthropic builds an “AI microscope” to probe Claude’s internal anatomy

How Alibaba built its compact but powerful video generation models. Towards a unified text-image diffusion model. A new approach to vision-language understanding from Alibaba. Microsoft adapts OpenAI models to build data workforce agents.

Students in a computer lab using AI tools to generate digital art, including rockets, landscapes, portraits, and 3D characters.

Data Points

Gemini 2.5 Pro takes the top spot on key benchmarks: GPT-4o’s popular but controversial new image generator

DeepSeek updates its V3 model with new skills and MIT license. Reve Image 1.0 excels at text and typography design. Qwen2.5-Omni tackles text. Images, audio, and video. Software developers have tried AI, but some like it better.

Visual model aligning diffusion embeddings with DINOv2 encoders using REPA and DiT/SiT blocks.

Machine Learning Research

Faster Learning for Diffusion Models: Pretrained embeddings accelerate diffusion transformers’ learning

Diffusion transformers learn faster when they can look at embeddings generated by a pretrained model like DINOv2.

AI tutoring system interface showing real-time context integration, privacy, and expert-like feedback generation.

Tech & Society

LLM Support for Tutors: GPT-4 boosts remote tutors’ performance in real time, study finds

Students benefit from tutoring, but training tutors is expensive. A study shows that large language models can boost tutors’ effectiveness in real time.

Diagram comparing diffusion, flow matching, and shortcut models for image generation with fewer steps.

Machine Learning Research

Better Images in Fewer Steps: Researchers introduce shortcut models to speed up diffusion

Diffusion models usually take many noise-removal steps to produce an image, which takes time at inference. There are ways to reduce the number of steps, but the resulting systems are less effective. Researchers devised a streamlined approach that doesn’t sacrifice output quality.

Comparison table of Gemini and Gemma models across benchmarks like MMLU, MATH, and GPQA with radar charts.

Machine Learning Research

Vision-Language, Compact and Open: Google releases Gemma 3 vision-language models with open weights

Google updated its open-weights family of large language models to include versions that handle image and video inputs.

Cartoon of a man playing violin saying “I’m fine-tuning!” while a woman at her desk covers her ears, replying “Did you try prompting?”

Letters

When to Fine-Tune — and When Not To: Many teams that fine-tune their models would be better off prompting or using agentic workflows. Here's how to decide.

Fine-tuning small language models has been gaining traction over the past half year.

The Batch Newsletter

Compact Vision-Language with Open Weights, Faster Learning, Diffusion in Few Steps, LLMs Aid Tutors

The Batch AI News and Insights: Fine-tuning small language models has been gaining traction over the past half year.

Aerial view of a hedge maze with a large black spider at the center, surrounded by trees, benches, and paths.

Data Points

Building a model for vision and speech: How Cloudflare thwarts unauthorized AI crawlers… by using AI

Nvidia’s Nemotron adds reasoning to Llama models. Does ChatGPT make frequent users more lonely? OpenAI’s o1-pro costs a pretty penny. Mistral Small 3.1 gives Gemma 3 27B some competition.

Children learning programming in a modern classroom with laptops, robots, and screens displaying code, promoting STEM education.

Data Points

This Aardvark predicts the weather: GPT-4o meets Whisper; OpenAI’s new models

Nvidia gives Project DIGITS a new name. AI models compete to build Minecraft items. Claude chatbot now includes search. A Moore’s law-like regularity for AI agents.

AYA Vision architecture diagram showing vision encoder, multimodal merging, and LLM backbone for AI-powered image processing.

The Batch Newsletter

Inside Google’s Co-Scientist, Copyright Office Weighs Generated Works, Multilingual (and Good at All of Them), Diffusion for Materials Design

The Batch AI News and Insights: Last Friday on Pi Day, we held AI Dev 25, a new conference for AI Developers.