ID 原文 译文
46576 实验表明引入 Q-Learning 进行码率调整的自适应算法在码率切换的稳定性方面表现较好。 The experiments were employed by the Q-learning rate control approach in two typical HAS algorithms. The experiments show the rate control approach can enhance the stability of rate switching in HAS clients.
46577 为解决 Hadoop 云平台无法动态控制用户访问请求的问题，提出一种基于用户行为评估的 Hadoop 云平台动态访问控制（DACUBA, dynamic access control based on user behavior assessment）模型。 In order to solve the problem that Hadoop cloud platform could not dynamically control user access request, a Hadoop cloud dynamic access control model based on user behavior assessment (DACUBA) was proposed.
46578 该模型首先实时收集用户指令序列，通过并行指令序列学习（PCSL, parallel command sequence learning）获取用户行为轮廓。 The model first collected the user instruction sequence in real time and the user behavior contour was obtained by parallel command sequence learning (PCSL).
46579 然后利用前向轮廓建立全局 K 模型，对后续行为序列进行分类并对分类结果进行评估。 Then the global K model was established by using the forward profile, the subsequent sequence was classified and the classification results were evaluated.
46580 随后将评估结果与改进 Hadoop 访问控制机制结合，使云平台用户的访问权限随自身行为动态改变。 The evaluation results were combined with the improved Hadoop access control mechanism to make the cloud platform users’ access rights change dynamically with their own behaviors.
46581 最后通过实验验证了模型算法的有效性和动态访问控制机制的可行性。 Experimental results demonstrate that the model algorithm is effective and the dynamic access control mechanism is feasible.
46582 以往的自适应 JPEG 隐写算法大多只在秘密信息嵌入之前计算图像失真，无法在嵌入过程中动态地调节失真代价。 Previous adaptive JPEG steganography algorithms mostly calculate image distortion before secret information embedding, so they can’t dynamically adjust distortion costs.
46583 考虑到信息嵌入时的交互影响，提出一种动态更新失真代价的自适应 JPEG 隐写算法。 Considering the mutual impacts during embedding, an adaptive JPEG steganography algorithm based on distortion cost updating was proposed.
46584 首先，分析量化步长、待嵌 DCT(discrete cosine transform)系数绝对值以及扰动误差 3 种影响嵌入波动的因素。 Firstly, three factors that affect embedding fluctuations were analyzed, including quantization step, the absolute value of the quantized DCT coefficient to be modified and perturbation error.
46585 进而，提出失真代价更新策略 EUS(embedding update strategy)，使图像的失真代价能够动态地更新。 Then the embedding update strategy (EUS) was proposed, which enabled to dynamically update the distortion costs.