Jiarui Feng

Mixture-of-LoRAs models can finally leverage their full potential: ReMix solves the long-standing problem of imbalanced LoRA usage by using reinforcement learning to train a router with non-learnable weights, leading to significant performance gains.

Generating realistic tabular data with both numbers and free-form text just got easier: TabDLM bridges the gap between diffusion models and LLMs for superior joint modeling.