About
Research & Development
Unpacking Text Complexity: How Simplified Data Shapes Language Model Learning
TLDR: This research explores how text complexity in pretraining data affects language models. It finds that while simplifying text doesn't significantly hurt general language...
Research & Development
Beyond Words: Evaluating AI’s Grasp of Emotion and Context in Speech
TLDR: CP-Bench is a new benchmark for speech-LLMs that evaluates their ability to understand both verbal content and non-verbal cues like emotion and prosody...
Research & Development
PruneCD: Enhancing LLM Factuality Through Intelligent Layer Pruning
TLDR: PruneCD is a new method that improves the factual accuracy of Large Language Models (LLMs) by using a technique called contrastive decoding. Unlike...
Research & Development
FakeChain: Uncovering the Weaknesses in Multi-Step Deepfake Detection
TLDR: A new research paper introduces FakeChain, a benchmark for multi-step deepfakes, revealing that current detection models are heavily biased towards the final manipulation...
Research & Development
Diffusion LLMs Enhance Speech Recognition and Transcript Refinement
TLDR: A study introduces Whisper-LLaDA, an ASR system integrating a Whisper encoder with the LLaDA diffusion LLM. It significantly improves speech recognition accuracy through...
Research & Development
Bridging the Performance Gap: How Small AI Models Learn from Large Ones Without Labeled Data
TLDR: The Model Parity Aligner (MPA) is a new framework that significantly improves the performance of Small Vision-Language Models (S-VLMs) in visual question answering...