AI Research Today
autor: Aaron
How Does a Diffusion Model Work, Part 2: Flow Matching and Diffusion
O tym podcaście
AI Research Today unpacks the latest advancements in artificial intelligence, one paper at a time. We go beyond abstracts and headlines, walking through architectures, experiments, training details, ablations, failure modes, and the implications for future work. Each episode will choose between one and three new, impactful research papers and go through them in depth. We will discuss the papers at the level of an industry practitioner or AI researcher. If you want to understand the newest topics in AI research but don't have the time to dig through the papers yourself, this is your solution.
Odcinki
- How Does a Diffusion Model Work, Part 2: Flow Matching and Diffusion · 37 min · Sezon 1 Odcinek 14Send us Fan Mail Diffusion models have become the foundation of modern generative AI, powering state-of-the-art systems for image generation, video synthesis, protein design, and…
- How Does a Diffusion Model Work, Part 1: Introduction · 22 min · Sezon 1 Odcinek 13Send us Fan Mail Diffusion models have become the foundation of modern generative AI, powering state-of-the-art systems for image generation, video synthesis, protein design, and…
- Generative Recursive Reasoning · 37 min · Sezon 1 Odcinek 12Send us Fan Mail In this episode, we explore the paper "Generative Recursive Reasoning (GRAM)," a fascinating new approach to AI reasoning co-authored by Yoshua Bengio and…
- OccuBench: Evaluating AI Agents on Real-World Professional Tasks via Language Environment Simulation · 33 min · Sezon 1 Odcinek 10Send us Fan Mail In this episode, we break down the new paper “OccuBench: Evaluating AI Agents on Real-World Professional Tasks via Language Environment Simulation,” which…
- GradMem: Teaching LLMs to Remember (Without Retraining) · 29 minSend us Fan Mail In this episode, we break down GradMem, a new approach to memory in large language models: https://arxiv.org/pdf/2603.13875v1 Instead of relying on the…
- Language Models are Injective and Hence Invertible · 27 minSend us Fan Mail In this episode, we break down a fascinating new result from recent research: that modern Transformer language models are almost surely injective—meaning…
- Learning to Reason in 13 Parameters · 27 min · Sezon 1 Odcinek 7Send us Fan Mail Link to arxiv: https://arxiv.org/pdf/2602.04118 Large language models have recently shown impressive reasoning abilities, often learned through reinforcement…
- SPIRAL: Symbolic LLM Planning via Grounded and Reflective Search · 29 min · Sezon 1 Odcinek 6Send us Fan Mail Large Language Models often struggle with complex planning tasks that require exploration, backtracking, and self-correction. Once an LLM commits to an early…
- Meta-RL Induces Exploration In Language Agents · 29 min · Sezon 1 Odcinek 5Send us Fan Mail Episode Paper: https://arxiv.org/pdf/2512.16848 In this episode, we dive into a cutting-edge AI research breakthrough that tackles one of the biggest challenges…
- DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search · 37 min · Sezon 1 Odcinek 4Send us Fan Mail In this episode, we unpack DeepSearch, a new paradigm in reinforcement learning with verifiable rewards (RLVR) that aims to overcome one of the biggest…
- Transformer-Squared: Self-Adaptive LLMs · 40 min · Sezon 1 Odcinek 3Send us Fan Mail In this episode we’re diving into “Transformer-Squared: Self-Adaptive LLMs” — a new framework for adapting large language models to unseen tasks on the fly by…
- Nested Learning: The Illusion of Deep Learning Architectures · 50 min · Sezon 1 Odcinek 2Send us Fan Mail NL.pdf In this episode, we dive into Nested Learning (NL) — a new framework that rethinks how neural networks learn, store information, and even modify…
- AgentEvolver: An Autonomous Agent Framework · 42 min · Sezon 1 Odcinek 1Send us Fan Mail https://arxiv.org/pdf/2511.10395 What if AI agents could teach themselves? In this episode, we dive into AgentEvolver, a groundbreaking framework from Alibaba's…








