LossSeq
defimported Lean
A loss sequence $\ell$ for $N$ experts over $T$ rounds is a function $\ell : \mathrm{Fin}\,T \to \mathrm{Fin}\,N \to \mathbb{R}$, i.e., at each round $t < T$ the adversary reveals a loss vector $\ell_t : \mathrm{Fin}\,N \to \mathbb{R}$ whose values are intended to lie in $[0,1]$.
Statement
set_option relaxedAutoImplicit false in set_option autoImplicit false in open Real Finset BigOperators in def LossSeq (N T : ℕ)
- Defined in
- TCSlib, `LossSeq`
- Lean proof
- By Karim Abdel Sadek, Mark Bedaywi, from github.com/Shilun-Allan-Li/tcslib at
c8a02591b1e5(Copyright (c) 2026 Karim Abdel Sadek and Mark Bedaywi. All rights reserved, Apache-2.0). Compiled here against leanprover/lean4:v4.33.1, Mathlib v4.33.1; written for leanprover/lean4:v4.25.0.LossSeq← LossSeq (TCSlib/LearningTheory/Hedge/Basic.lean:69): verbatim
- Verified
- 07 Oct 2026
- Axioms
- none listed
- Used by 29
- hedge_regret_bound_tightTight Hedge regret bound. Fix $N \ge 1$ experts and a horizon of $T$ rounds, and run the Hedge algorithm with a learning rate $\eta > 0$.…
- cumLoss_horizonCumulative loss at the horizon. Fix $N$ experts and a horizon $T$, and let $\ell$ be a loss sequence assigning to each round $t$ and…
- hedgeWeightThe unnormalized Hedge weight of expert $i$ at time $t$ with learning rate $\eta$ is…
- log_potential_step_tightTight per-step logarithmic potential bound for Hedge. Fix $N \ge 1$ experts and a learning rate $\eta > 0$, and let $\ell$ be a valid loss…
- hedgeDistThe Hedge distribution at round $t$ is the normalization of the weight vector: \[ p_t(i) \;=\; \frac{w_t(i)}{W_t}. \]
- potential_ratio_lePotential ratio bound for Hedge. Fix $N \ge 1$ experts, a horizon of $T$ rounds, a learning rate $\eta > 0$, and a valid loss sequence…
- hedgeWeight_succRecursion for the Hedge weight. Fix a learning rate $\eta \in \bbr$ and a loss sequence $\ell$ for $N$ experts over $T$ rounds. For every…
- potential_posPositivity of the Hedge potential. Fix a learning rate $\eta \in \bbr$, a loss sequence $\ell$ for $N \ge 1$ experts over $T$ rounds, and…
- ZeroSumGame.toLossSeqGiven a game $G$ and a sequence of column responses $j_0, \dots, j_{T-1}$, the induced loss sequence for the row player is…
- hedge_constructionHedge yields an approximate minimax pair. Let $G$ be a finite two-player zero-sum game with $M \ge 2$ row actions and $N \ge 1$ column…
- hedgeWeight_posPositivity of Hedge weights. Fix a learning rate $\eta \in \bbr$, a loss sequence $\ell$ for $N$ experts over $T$ rounds, a time $t$, and…
- LossSeq.ValidA loss sequence $\ell$ is valid if all individual losses are in $[0,1]$: $\forall t\,i,\; 0 \le \ell_t(i) \le 1$.
- log_potential_telescopeTelescoping of the log potential. Let $\eta \in \bbr$ and let $\ell$ be a loss sequence for $N$ experts over $T$ rounds, with potential…
- hedgeLoss_le_oneHedge's expected loss is at most one. Fix a learning rate $\eta \in \bbr$, and let $\ell$ be a valid loss sequence for $N \ge 1$ experts…
- hedgeCumLossThe cumulative loss of the Hedge algorithm over all $T$ rounds is \[ \widehat{L}_T \;=\; \sum_{t=0}^{T-1} \widehat{\ell}_t. \]
- hedgeLoss_nonnegNonnegativity of Hedge's expected loss. Let $\ell$ be a valid loss sequence for $N \ge 1$ experts over $T$ rounds, fix any learning rate…
- hedgeLossThe expected loss of the learner at round $t$ under the Hedge distribution is…
- log_potential_stepOne-step logarithmic potential bound for Hedge. Fix $N \ge 1$ experts and $T$ rounds, a learning rate $\eta > 0$, and a valid loss…
- regretThe regret of the Hedge algorithm is the excess cumulative loss over the best expert in hindsight:…
- cumLossThe cumulative loss of expert $i$ through the first $t$ rounds is \[ L_t(i) \;=\; \sum_{s < t} \ell_s(i), \] where the sum ranges…
- potentialThe potential (sum of unnormalized weights) at time $t$ is \[ W_t \;=\; \sum_{i=1}^{N} w_t(i). \] The potential telescopes to…
- regret_le_of_log_potential_stepRegret from a per-step log-potential bound. Let $N \ge 1$, let $\eta > 0$, let $\ell$ be a loss sequence for $N$ experts over $T$ rounds,…
- hedgeDist_nonnegNonnegativity of the Hedge distribution. Fix $N \ge 1$ experts and $T$ rounds, a learning rate $\eta \in \bbr$, and a loss sequence…
- bestExpertLossThe best expert loss in hindsight is the minimum cumulative loss achieved by any single expert:…
- potential_zeroInitial value of the potential. Fix $N \ge 1$ experts, a horizon $T$, and a loss sequence $\ell$ over these experts and rounds, and let…
- cumLoss_succCumulative loss recurrence. Let $\ell$ be a loss sequence for $N$ experts over $T$ rounds, and write $L_t(i) = \sum_{s < t} \ell_s(i)$ for…
- hedgeDist_sumThe Hedge distribution sums to one. Fix a number of experts $N \ge 1$, a horizon $T$, a learning rate $\eta \in \bbr$, a loss sequence…
- potential_ge_best_expertPotential lower bound via the best expert. Fix $N \ge 1$ experts and $T$ rounds, a loss sequence $\ell$, and a learning rate $\eta > 0$.…
- hedge_regret_boundRegret bound for the Hedge algorithm. Fix $N \ge 1$ experts and $T$ rounds, and let $\ell$ be a valid loss sequence, so that…
Lean source view module on GitHub
set_option relaxedAutoImplicit false in set_option autoImplicit false in open Real Finset BigOperators in /-- The type of loss sequences over `N` experts and `T` rounds: a whole `T × N` trajectory, `ℓ t i` being the loss of expert `i` at round `t` [CBL06, §2.1]. The `[0, 1]` range condition is kept as a separate predicate (`LossSeq.Valid`) instead of being built into the type, which keeps the algebraic definitions simple. -/ def LossSeq (N T : ℕ) := Fin T → Fin N → ℝ