第 20 / 31 页
非代码面试题
显示 20 / 612 道匹配题目
答题状态：未尝试未正确已正确
ID题目领域难度题型进度权限
3193Total Cost Over a Negative-Binomial HorizonLet X 1,X 2,\dots be i.i.d. increments with E[X i]=4 and Var (X i)=6. Let N be independent of the increments and distributed as NegativeBinomial(r=3, p= 2 5 ). For the stopped sum S N=\sum i=1 N X i, compute E[S N] and Var (S N).概率中等derivation未尝试面试订阅3201Expected Trials to Reach 5 SuccessesIndependent Bernoulli trials succeed with probability 2 5 . Let T be the first time the cumulative number of successes reaches 5. Use Wald-style reasoning to compute E[T].概率中等derivation未尝试面试订阅3206Variance of Trials to Reach 5 SuccessesIndependent Bernoulli trials succeed with probability 2 5 . Let T be the first time the cumulative number of successes reaches 5. Use Wald-style second-moment reasoning to compute Var (T).概率困难derivation未尝试面试订阅3212Second Moment of Centered Sum at a Poisson HorizonLet X 1,X 2,\dots be i.i.d. with mean and variance 3. Let N be independent of the increments and distributed as Poisson(4). Show that for the centered stopped sum M N=\sum i=1 N (X i- ), one has E[M N 2] equal to what value?概率中等derivation未尝试面试订阅3214Centered Slippage Variance Under Random StoppingLet X 1,X 2,\dots be i.i.d. with mean and variance 4. Let N be independent of the increments and distributed as Geometric( 1 3 ). Show that for the centered stopped sum M N=\sum i=1 N (X i- ), one has E[M N 2] equal to what value?概率中等derivation未尝试面试订阅3218Small Sample, One Success: Which Lens Shrinks More Naturally?A new quoting rule is tried 5 times and succeeds once. Why does a Bayesian analysis with a skeptical prior naturally shrink the estimated success probability toward a baseline, while a plain frequentist point estimate does not do that unless you add some regularization device on top?统计中等essay未尝试面试订阅3224Hierarchical Bayes Versus BonferroniYou are screening 200 alphas and most are probably zero. Why does hierarchical Bayes approach this problem differently from Bonferroni-style frequentist correction?统计中等essay未尝试面试订阅3227Sequential Updates Without Design LockData trickle in continuously and the desk wants to update beliefs every hour. Why is Bayesian inference naturally sequential, while a frequentist testing workflow often needs more design discipline to preserve its advertised guarantees?统计中等essay未尝试面试订阅3230Sparse Event Rates and Online CalibrationWhy is a Bayesian approach often attractive when you are updating a very sparse event rate online, such as rare failures or rare fills?统计中等essay未尝试面试订阅3234Partial Pooling Versus Separate FitsWhy does a hierarchical Bayesian model for many related assets often produce more stable estimates than fitting each asset separately and then testing them one by one?统计简单essay未尝试面试订阅3426Remaining Entropy After Revealing a 3-4-5 Bucket LabelA source is uniform over 12 states. You reveal which bucket the state lies in, where the bucket sizes are 3, 4, and 5. What is the remaining entropy H(X|bucket)?数学简单derivation未尝试面试订阅3427Remaining Entropy After Revealing a 1-3-4-8 PartitionA source is uniform over 16 states. A side signal reveals whether the realized state lies in a block of size 1, 3, 4, or 8. What is H(X|signal)?数学简单derivation未尝试面试订阅3428Remaining Entropy After a 4-vs-6 Split SignalA source is uniform over 10 states. You reveal whether the outcome lies in the first 4 states or in the last 6 states. What is the remaining entropy?数学简单derivation未尝试面试订阅3429Entropy of a Three-Regime Disjoint Alphabet SourceA regime label takes values A, B, C with probabilities 0.2, 0.5, 0.3. Conditional on A the source is deterministic, conditional on B it is uniform over 4 symbols, and conditional on C it is uniform over 2 symbols. Assuming the symbol sets are disjoint across regimes, what is the total source entropy?数学简单derivation未尝试面试订阅3430Entropy of a Fair 3-vs-5 Regime MixtureA fair regime bit first chooses source A or B. Source A is uniform over 3 symbols and source B is uniform over 5 symbols, with disjoint alphabets. What is the total entropy of the emitted symbol?数学简单derivation未尝试面试订阅3434Remaining Entropy After Revealing 'Special Pair or Not'A source is uniform over 8 states. A side signal reveals whether the state lies in a special pair of states or in the other 6 states. What is the remaining entropy?数学中等derivation未尝试面试订阅3435Entropy Reduction From Revealing One Prefix BitA source is uniform over 16 equiprobable states. If a side signal reveals the first binary prefix bit of the state label, by how many bits does entropy drop?数学中等derivation未尝试面试订阅3436Why Uniform Distributions Maximize Entropy Under a Support ConstraintWhy does spreading probability mass more evenly over a fixed finite support raise entropy?数学中等essay未尝试面试订阅3437Why Coarse-Graining Lowers EntropyWhy does merging labels generally reduce entropy rather than increase it?数学中等essay未尝试面试订阅3438Why Side Information Cannot Increase Remaining UncertaintyWhy is it impossible for an informative side signal to make the conditional entropy of the original source larger on average?数学中等essay未尝试面试订阅