[论文解读] Verification of deep probabilistic models
本文提出了一种用于深度概率模型(如神经过程和变分自编码器)的新验证框架,确保在所有条件输入下,通过对潜在变量采样,模型输出以高概率满足线性约束。该方法利用凸对偶性和误差函数的概率界,高效计算合规概率的严格下界,成功验证了函数回归任务中预期的性质,如单调性和凸性。
Probabilistic models are a critical part of the modern deep learning toolbox - ranging from generative models (VAEs, GANs), sequence to sequence models used in machine translation and speech processing to models over functional spaces (conditional neural processes, neural processes). Given the size and complexity of these models, safely deploying them in applications requires the development of tools to analyze their behavior rigorously and provide some guarantees that these models are consistent with a list of desirable properties or specifications. For example, a machine translation model should produce semantically equivalent outputs for innocuous changes in the input to the model. A functional regression model that is learning a distribution over monotonic functions should predict a larger value at a larger input. Verification of these properties requires a new framework that goes beyond notions of verification studied in deterministic feedforward networks, since requiring worst-case guarantees in probabilistic models is likely to produce conservative or vacuous results. We propose a novel formulation of verification for deep probabilistic models that take in conditioning inputs and sample latent variables in the course of producing an output: We require that the output of the model satisfies a linear constraint with high probability over the sampling of latent variables and for every choice of conditioning input to the model. We show that rigorous lower bounds on the probability that the constraint is satisfied can be obtained efficiently. Experiments with neural processes show that several properties of interest while modeling functional spaces can be modeled within this framework (monotonicity, convexity) and verified efficiently using our algorithms
研究动机与目标
- 开发一种形式化验证框架,用于深度概率模型,确保在潜在变量采样下,模型输出以高概率满足期望属性。
- 解决概率模型中 worst-case 验证的局限性,后者通常产生空洞或过度保守的结果。
- 实现对神经过程和条件 VAE 等模型中单调性、凸性和语义不变性等属性的高效、可扩展验证。
- 即使在潜在变量存在随机性的情况下,也提供模型输出满足线性约束的概率的计算可行下界。
提出的方法
- 将验证问题表述为:对于集合 X 中的所有条件输入 x,满足 cᵀf(x,z) + d ≥ 0 的概率需被限制在 ε 以内,即 P(cᵀf(x,z) + d ≥ 0) ≤ ε。
- 利用凸对偶性推导出上界线性形式 cᵀf(x,z) + d 的对偶函数 G(λ, z),实现概率松弛。
- 应用弱对偶性,通过 P(G(λ, z) ≥ 0) + P(z ∉ [α, β]) 来界定违反概率,其中 [α, β] 是潜在空间的有界区域。
- 利用 z 的高斯分布,通过互补误差函数(erfc)以闭式表达概率 P(G(λ, z) ≥ 0)。
- 通过梯度下降优化对偶变量 λ 和边界 α, β,参数化 β = α + η² 以实现无约束优化。
- 采用重参数化方法处理非标准正态潜在变量,通过变换将其转化为标准正态变量。
实验结果
研究问题
- RQ1我们能否验证深度概率模型在所有条件输入下,以高概率产生满足线性约束的输出?
- RQ2如何高效计算模型输出满足给定规范(如单调性或凸性)的概率的下界?
- RQ3所提出的框架能否应用于神经过程等函数回归模型,以验证预测函数的结构属性?
- RQ4即使潜在变量从高斯分布中采样,该方法在实践中是否仍能产生非空洞、非保守的边界?
主要发现
- 该验证框架成功计算出模型输出满足线性约束的概率的非空洞、严格下界,即使在高维潜在空间中亦成立。
- 对于在累积分布函数上训练的神经过程,该方法验证了预测输出在目标输入区间内保持在预期范围内(如 [0,1])的概率 ≥ 0.99。
- 预测值的上下界随输入 δ 单调递增,确认在整个定义域内呈现预期的 CDF 类行为。
- 尽管问题本身非凸,但通过在 λ、α 和 β 上进行梯度下降优化概率边界,在实践中收敛迅速。
- 该方法证明了神经过程可被验证以生成与已知函数属性(如单调性和有界性)一致的输出。
更好的研究,从现在开始
从阅读论文到最终审阅,大幅缩短您的研究时间。
无需绑定信用卡
本解读由 AI 生成,并经人工编辑审核。