用 Counter-Strike 2 遥测数据建模 60-120ms 快速人类反应
Toward Fast Human Reaction Modeling from High-Frequency Digital Interaction Data
研究者拿 Counter-Strike 2 的毫秒级对局数据训练模型预测玩家 60-120ms 内的本能反应，EMBER 准确率 73.5%，还比 LLM 基线快好几个量级，思路挺有意思。
一项 arXiv 探索性研究利用 Counter-Strike 2 的毫秒级遥测数据，构建了 Ego-Knowledge 感知状态重建管线，在 60-120ms 窗口内从可视和受击事件提取战或逃标签。团队训练了两个遥测原生模型：梯度提升模型 EMBER 达到 73.5% 准确率，自研 transformer ORBIT 达 71.2%，均显著高于 51.5% 的多数类基线。在 1,000 个验证样本上，最强 LLM 基线仅 62.7%，EMBER 以 73.6% 胜出（McNemar p=1.3e-9）且推理成本和延迟低几个数量级。加入触发后短暂窗口后，准确率单调提升至 85.15%，作者提出该曲线可作为快速控制向审议控制转变的候选探针，但强调神经科学解释仍是假设生成性质。
Toward Fast Human Reaction Modeling from High-Frequency Digital Interaction Data
We investigate whether limbic-like signatures of rapid, salience-driven behavior can be identified in high-resolution Counter-Strike 2 telemetry. This exploratory study introduces Ego-Knowledge, a leak-reduced reconstruction of each player's plausible perceptual state, supported by a millisecond-precision visibility engine and a reaction-centered extraction pipeline. Using a neuroscience-motivated 60--120 ms window, we derive coarse fight-or-flight labels from visibility- and damage-triggered events, with back-damage cases enabling constrained reaction-time analysis and manual label checks. We train two telemetry-native models using only information available by the trigger: EMBER (gradient boosting) and ORBIT (a custom transformer). On the full validation set, EMBER achieves 73.5% accuracy and ORBIT 71.2%, versus a 51.5% majority-class baseline, and both generalize to held-out players. The strongest LLM baseline achieves 62.7% on 1,000 validation cases; on these same cases, EMBER (73.6%) and ORBIT (70.9%) perform better (McNemar p=1.3e-9 and 1.5e-6, respectively), with orders-of-magnitude lower inference time and cost, though the comparison is not compute-matched. Including a short post-trigger window changes forecasting into early response recognition, raising accuracy monotonically to 85.15%. We propose this curve as a candidate probe of the transition from rapid to deliberative control, but roughly uniform per-tick gains do not establish that interpretation. The neuroscience framing is hypothesis-generating, not mechanistic: physiological ground truth, suitable control distributions, and reverse-inference limitations remain open. We outline extensions to auditory reconstruction, larger datasets, and telemetry-native foundation models.