Brief

Aflac general counsel: Georgia lawmakers took a crucial step forward on sickle cell disease – but there’s more work to be done

Georgia has passed the Sickle Cell Disease Protection Act, requiring annual reviews of Medicaid coverage for sickle cell disease treatments and services. This legislation, signed into law by Governor Brian Kemp, follows similar laws in Louisiana, Virginia, and Tennessee. While federal funding exists for research and treatment, it doesn't guarantee nationwide access or mandate state Medicaid program updates for new therapies. The act aims to ensure that medical breakthroughs for sickle cell disease translate into better real-world outcomes for patients. AI

• Georgia
• Sickle Cell Disease Protection Act
• Aflac
• Brian Kemp
• Louisiana
• Virginia
• Tennessee
• United States Department of Health and Human Services
• Sickle Cell Disease
• Aflac Cancer and Blood Disorders Center of Children’s Healthcare of Atlanta

Speculative Coupled Decoding for Training-Free Lossless Acceleration of Autoregressive Visual Generation

Researchers have introduced Variable Codebook Size Quantization (VCQ) to address limitations in autoregressive visual generation models. VCQ modifies the codebook size dynamically along the sequence, improving reconstruction performance and reducing the gFID score significantly on datasets like ImageNet. Additionally, new methods like VVS and Speculative Coupled Decoding (SCD) are accelerating inference speeds for these models by optimizing speculative decoding techniques, reducing the number of forward passes required while maintaining generation quality. AI

IMPACT These advancements in quantization and speculative decoding promise faster and more efficient visual generation models, potentially lowering inference costs and enabling new applications.

• gFID
• Transformer
• arXiv
• ImageNet

Rule2DRC: Benchmarking LLM Agents for DRC Script Synthesis with Execution-Guided Test Generation

Researchers are developing new methods to improve the evaluation and training of large language models (LLMs). One approach, SCOPE, calibrates LLM judges to ensure reliable pairwise evaluations with controlled error rates. Another technique, D3, uses dynamic influence graphs to optimize data scheduling during LLM training by considering sample interactions. Additionally, OBCache offers a principled framework for pruning key-value caches to reduce memory overhead during long-context inference, improving accuracy. AI

IMPACT New research introduces methods for more reliable LLM evaluation, efficient training data scheduling, and optimized inference, potentially improving LLM performance and resource utilization.

• LLMs
• PagedAttention
• FlashAttention
• A100 GPU
• LLM
• Nested WAIT
• Llama-2-7B
• Asteria
• vLLM
• KVDrive
• Orca
• Sarathi-Serve
• SCICONVBENCH
• A100
• FasterTransformer
• TIDE
• LLaDA2.0-mini
• LLaDA2.0-flash
• POPE benchmark
• V* benchmark
• LLMEval-Logic
• DeepSeek-R1-Distill-7B
• FT-Dojo
• FT-Agent
• rePIRL
• PALS
• LlamaWeb
• Charon
• WebGPU
• arXiv
• Frontier
• llama.cpp
• OBCache
• SCOPE
• Hermes
• Lean
• Item Response Theory
• LoRA
• AxBench
• Qwen
• LLaMA
• GPT-5
• Gemini 3 Pro
• FEM-Bench