Better Calibration, Utility

The ACUTE Protocol: Operationalizing Language Model Activations for Better Calibration, Utility, and Trust

new Abstract: As language models improve and become increasingly deployed to solve a variety of tasks, trustworthiness becomes essential. Calibration is a good proxy for trust: well-calibrated confidence estimates help inform the risk versus reward tradeoff when trusting a specific model output. Unfortunately, even as models improve, they remain poorly calibrated, often biasing towards overconfidence.

AlphaQ: Calibration-Free Bit Allocation for Mixture-of-Experts Quantization

Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures scale model capacity through sparse expert activation, but their deployment remains memory-bound because all expert weights must reside in memory. Mixed-precision quantization can substantially reduce this footprint by assigning different bit-widths to different experts. Existing approaches, however, typically rely on calibration data to estimate expert importance and determine bit allocation.

Dynamic Gradient-Based Calibration for Robust and Accurate Traffic Macrosimulation

arXiv:2605.19056v2 Announce Type: replace Abstract: Robust and accurate calibration of macroscopic traffic flow models such as METANET is critical for reliable prediction and effective control. While gradient-based methods are desirable for high-dimensional parameter spaces, their application to real-world traffic scenarios is hindered by highly nonconvex optimization landscapes. Consequently, standard static calibration frequently yields parameter sets that produce unstable, unrealistic...

Ask HN: What are tools you have made for yourself since the advent of AI?

I've made a number of ceramic molds for slumping fused glass into bowls. As well as wooden templates for ceramic mugs. I've devised a few carrying tools to move glass frit paintings from my studio down to my barn where the kilns sit without spilling the glass.

Why Everyone Hates AI Data Centers

KVarN: Native vLLM backend for KV-cache quantization by Huawei

⚡️ Built for agentic and long-context workloads. 💡 KVarN delivers 3-5x more KV-cache capacity and up to ~1.3x the throughput of FP16, so you fit far longer contexts and serve more concurrent requests, with FP16-level accuracy. 🔌 Calibration-free, plug-and-play with vLLM.

Cloud-tested quantum noise model predicts superconducting qubit errors with sevenfold better accuracy

Cloud-tested quantum noise model predicts superconducting qubit errors with sevenfold better accuracy Gaby Clark Scientific Editor Robert Egan Associate Editor Researchers from the Johns Hopkins Applied Physics Laboratory (APL) in Laurel, Maryland, and Johns Hopkins University in Baltimore have developed a practical, comprehensive noise-modeling framework for a popular class of superconducting quantum processors. Their work, published in the journal PRX Quantum, offers a sevenfold...

Deep learning four decades of human migration

Abstract Human migration is a fundamental driver of global demographic change, shaping population structure, labour markets and social policy across countries1,2,3. Although long-term migration patterns are often linked to economic development4, they can shift rapidly in response to shocks such as conflict, environmental crises and political change5. Despite its importance, migration remains difficult to measure consistently: existing data are sparse, concentrated in high-income settings and...