Paper Reviews by AI

ColorFlow, a new AI model, accurately colorizes black-and-white image sequences while preserving character identity.

Smaller is better: SLMs outperform LLMs in evolving complex & diverse instructions for AI training.

Training-free method adds physical properties to 3D models using vision-language models.

New benchmark for evaluating long-context models finds sub-O(n) methods lacking in real-world use cases.

New attack fools breast ultrasound AI using subtle text prompts.

From language models to action models: building AI that does things.

BLT: tokenizer-free LLM for efficiency and robustness

BrushEdit revolutionizes interactive image editing with instructions & inpainting.

Apollo LMMs achieve SOTA on video understanding tasks by exploring and optimizing the design and training of video-LMMs.

Word Sense Linking (WSL) revolutionizes word sense disambiguation by tackling its real-world limitations. It combines span identification and sense linking in plain text, offering better integration …

Norwegians show that using copyrighted material improves LLMs, but raises legal and ethical issues.

SynerGen-VL: A simpler, more powerful unified MLLM for image understanding and generation.

Shiksha: A new multilingual translation dataset and model surpasses existing benchmarks for Indian languages, focusing on scientific, technical, and educational domains.

RULEARENA, a new benchmark, rigorously evaluates large language models’ ability to apply complex, real-world rules across diverse scenarios, revealing significant shortcomings in current LLMs’ rule-gu…

Phi-4: a 14B parameter LLM surpassing its teacher model (GPT-4) in STEM-focused QA through innovative synthetic data generation and post-training techniques.

OLA-VLM boosts multimodal LLMs’ visual understanding by distilling knowledge from specialized visual encoders into the LLM’s internal representations during pretraining, achieving significant performa…

Neural LightRig uses multi-light diffusion to accurately estimate object normals and materials from a single image, outperforming existing methods.

VMB generates music from videos, images, and text, using description and retrieval bridges to improve quality and controllability.

Lyra: An efficient, speech-centric framework for omni-cognition, achieving state-of-the-art results across various modalities while being highly efficient.

LoRACLR merges multiple LoRA models for high-fidelity multi-concept image generation, using a contrastive objective to ensure concept distinctiveness and prevent interference.