Lean module · Foundations
BanditRLProof.Algorithms.HeavyTailAdaptive
Actual causal observations and their adaptive-count confidence.
Module map
Imports
BanditRLProof.HeavyTailArmLaw, BanditRLProof.HeavyTailGapThreshold
Imported by
Declarations
Open an item to read its exact compact statement and source link. Detailed teaching notes are linked when registered.
def
BanditRLProof.HeavyTail.robustMean
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robustMeanReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
noncomputable def robustMean (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (arm : Fin K) (t : ℕ) : ℝ
theorem
BanditRLProof.HeavyTail.robustMean_latent
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robustMean_latentReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robustMean_latent (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (arm : Fin K) (t : ℕ) : robustMean hK ε u stream arm t = (∑ s ∈ Finset.range (pullCount (robustAction hK ε u stream) arm t), truncate (sampleThreshold ε u t s) (stream s arm)) / pullCount (robustAction hK ε u stream) arm t
theorem
BanditRLProof.HeavyTail.robust_pullCount_pos
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robust_pullCount_posReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robust_pullCount_pos (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (arm : Fin K) (t : ℕ) (ht : K ≤ t) : 0 < pullCount (robustAction hK ε u stream) arm t
theorem
BanditRLProof.HeavyTail.robustMean_history
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robustMean_historyReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robustMean_history (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (arm : Fin K) (n : ℕ) : historyTruncatedMean (UCB.initializationArm hK 0) (sampleThreshold ε u (n+1)) n (ArmStreamPolicy.history (UCB.initializationArm hK 0) (nextArm hK ε u) stream n) arm = robustMean hK ε u stream arm (n+1)
theorem
BanditRLProof.HeavyTail.robustIndex_history
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robustIndex_historyReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robustIndex_history (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (arm : Fin K) (n : ℕ) : historyIndex (UCB.initializationArm hK 0) ε u n (ArmStreamPolicy.history (UCB.initializationArm hK 0) (nextArm hK ε u) stream n) arm = robustMean hK ε u stream arm (n+1) + confidenceRadius ε u (n+1) (pullCount (robustAction hK ε u stream) arm (n+1))
theorem
BanditRLProof.HeavyTail.robustMean_tail
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robustMean_tailReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robustMean_tail (hK : 0 < K) (ν : Kernel (Fin K) ℝ) [IsMarkovKernel ν] (arm : Fin K) (ε u : ℝ) (t : ℕ) (ht : K ≤ t) (hε0 : 0 ≤ ε) (hε : ε ≤ 1) (hu0 : 0 < u) (hX : Integrable (fun x : ℝ => x) (ν arm)) (hm : Integrable (fun x : ℝ => |x|^(1+ε)) (ν arm)) (hu : (∫ x, |x|^(1+ε) ∂ν arm) ≤ u) : (UCB.armStreamMeasure ν).real {stream | confidenceRadius ε u t (pullCount (robustAction hK ε u stream) arm t) ≤ |robustMean hK ε u stream arm t - ∫ x, x ∂ν arm|} ≤ t * (2 * Real.exp (-confidenceLog t))
theorem
BanditRLProof.HeavyTail.robust_selected_gap_le
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robust_selected_gap_leReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robust_selected_gap_le (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (mean : Fin K → ℝ) (best : Fin K) (n : ℕ) (hn : K ≤ n+1) (hbest : |robustMean hK ε u stream best (n+1) - mean best| ≤ confidenceRadius ε u (n+1) (pullCount (robustAction hK ε u stream) best (n+1))) (hchosen : |robustMean hK ε u stream (robustAction hK ε u stream (n+1)) (n+1) - mean (robustAction hK ε u stream (n+1))| ≤ confidenceRadius ε u (n+1) (pullCount (robustAction hK ε u stream) (robustAction hK ε u stream (n+1)) (n+1))) : mean best - mean (robustAction hK ε u stream (n+1)) ≤ 2 * confidenceRadius ε u (n+1) (pullCount (robustAction hK ε u stream) (robustAction hK ε u stream (n+1)) (n+1))
theorem
BanditRLProof.HeavyTail.robust_selected_small_radius_tail
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robust_selected_small_radius_tailReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robust_selected_small_radius_tail (hK : 0 < K) (ν : Kernel (Fin K) ℝ) [IsMarkovKernel ν] (best arm : Fin K) (ε u : ℝ) (t : ℕ) (ht : K ≤ t) (hε0 : 0 ≤ ε) (hε : ε ≤ 1) (hu0 : 0 < u) (hX : ∀ a, Integrable (fun x : ℝ => x) (ν a)) (hm : ∀ a, Integrable (fun x : ℝ => |x|^(1+ε)) (ν a)) (hu : ∀ a, (∫ x, |x|^(1+ε) ∂ν a) ≤ u) : (UCB.armStreamMeasure ν).real {stream | robustAction hK ε u stream t = arm ∧ 2 * confidenceRadius ε u t (pullCount (robustAction hK ε u stream) arm t) < (∫ x, x ∂ν best) - ∫ x, x ∂ν arm} ≤ 4*t*Real.exp (-confidenceLog t)
theorem
BanditRLProof.HeavyTail.robust_initial_count_zero
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robust_initial_count_zeroReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robust_initial_count_zero (hK : 0 < K) (ε u : ℝ) (stream : UCB.ArmRewardStream K) (t : ℕ) (ht : t < K) : pullCount (robustAction hK ε u stream) (robustAction hK ε u stream t) t = 0
theorem
BanditRLProof.HeavyTail.robust_large_count_tail
Compiled
No declaration docstring is present; use the chapter context and exact statement below.
Used in these reading views: Bandit Book
1. Finite bandits, traces, and regret
Canonical node identity
declaration:BanditRLProof.HeavyTail.robust_large_count_tailReading membership is not a proof dependency. Exact assumptions remain in the Lean statement.
theorem robust_large_count_tail (hK : 0 < K) (ν : Kernel (Fin K) ℝ) [IsMarkovKernel ν] (best arm : Fin K) (ε u : ℝ) (T t : ℕ) (ht : t ≤ T) (hε0 : 0 < ε) (hε : ε ≤ 1) (hu0 : 0 < u) (hgap : 0 < (∫ x, x ∂ν best) - ∫ x, x ∂ν arm) (hX : ∀ a, Integrable (fun x : ℝ => x) (ν a)) (hm : ∀ a, Integrable (fun x : ℝ => |x|^(1+ε)) (ν a)) (hu : ∀ a, (∫ x, |x|^(1+ε) ∂ν a) ≤ u) : (UCB.armStreamMeasure ν).real {stream | robustAction hK ε u stream t = arm ∧ gapThreshold ε u ((∫ x, x ∂ν best) - ∫ x, x ∂ν arm) T ≤ pullCount (robustAction hK ε u stream) arm t} ≤ 4*t*Real.exp (-confidenceLog t)