Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

Machine Learning

886 missions · 536 completed

The science of systems that learn from data and experience. Its scope runs from the statistical and mathematical foundations of learning, including generalization, expressivity, and computational limits, through the design of learning algorithms, deep learning, reinforcement learning, and probabilistic methods, to the empirical study of large models and the trustworthiness, interpretability, and societal impact of learned systems.

Missions

Open350Completed536All886
🏆Completed
ProbabilityStatisticsTheoretical Computer Science·Captain: mikedeng1

Foundations of Machine Learning I: The PAC Learning FrameworkTextbook

How many labeled examples does a learning algorithm need to see before its output generalizes well to unseen data? Chapter 2 of Foundations of Machine Learning answers this question for the simplest nontrivial setting — a finite hypothesis set — and in doing so introduces the book's central object, the Probably Approximately Correct (PAC) learning…

9 thms4 active users
🏆Completed
Convex OptimizationOperations ResearchOptimization·Captain: mikedeng1

First-Order and Stochastic Optimization Methods for Machine Learning VII: Gradient Sliding for Composite OptimizationTextbook

Composite convex programs — objectives split into a smooth piece and a nonsmooth piece — are ubiquitous in data analysis: LASSO-type inverse problems, regularized empirical-risk minimization, and total-variation-type image reconstruction all minimize f(x)+h(x)+χ(x)f(x)+h(x)+\chi(x)f(x)+h(x)+χ(x) over a convex set, where fff is smooth (a data-fidelity term, often expensive to…

5 thms4 active usersReviewed
🏆Completed
Convex OptimizationOperations ResearchOptimization·Captain: mikedeng1

First-Order and Stochastic Optimization Methods for Machine Learning III: Stochastic Mirror DescentTextbook

Machine learning's canonical training objective — minimize an expected or empirical risk over a data distribution — is almost never observed exactly: at each step an algorithm sees only a noisy gradient sample (a minibatch gradient, a single-example gradient, a simulation draw). Stochastic mirror descent (Nemirovski, Juditsky, Lan & Shapiro 2009) is the…

5 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization XIII: Blackwell's Approachability Theorem and Online Convex OptimizationTextbook

Von Neumann's minimax theorem (Chapter VIII) settles two-player zero-sum games with scalar payoffs. In 1956, Blackwell asked the natural generalization: what can a player guarantee in a repeated game with vector-valued payoffs, where "winning" means driving the average payoff into a target set rather than above a target value? For decades the resulting…

12 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization XII: The Online Boosting MethodTextbook

Chapter XI boosted a weak learner into a strong one for a single offline fit to a fixed sample. Chapter 12 asks the analogous question online: when the pool of experts is too large to run Hedge over directly (the contextual-learning setting, where "experts" are policies mapping contexts to actions and their number is exponential), can black-box access to a…

12 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization IX: From Online Convex Optimization to PAC LearningTextbook

Every algorithm in Chapters I–VIII minimizes regret, an online, adversarial performance measure with no reference to a data-generating distribution. Chapter 9 asks what regret minimization buys in the classical statistical learning setting, where examples are drawn i.i.d. from a fixed distribution and the goal is a hypothesis that generalizes well to unseen…

7 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization VII: The Online Conditional Gradient AlgorithmTextbook

Every algorithm through Chapter VI updates its iterate by a Euclidean projection onto the decision set KKK. For many decision sets that arise in practice — bounded-nuclear-norm matrices (matrix completion / recommendation systems), the flow polytope (network routing), the Birkhoff–von Neumann polytope (ranking/permutations), matroid polytopes — a projection…

10 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization IV: The Online Newton Step AlgorithmTextbook

Online convex optimization measures a decision maker against the best fixed decision in hindsight, and the standard guarantee — achieved, for instance, by online gradient descent — is regret growing like O(T)O(\sqrt T)O(T​) over TTT rounds. This rate is unimprovable for general convex losses: an adversary can always force Ω(T)\Omega(\sqrt T)Ω(T​) regret against any algorithm.…

13 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization X: Efficient Adaptive Regret for Online Convex OptimizationTextbook

Every regret guarantee through Chapter IX compares the algorithm to the single best fixed decision in hindsight. That comparison is meaningless when the environment itself changes: a commuter's best route differs on weekdays versus weekends, an investor's best portfolio differs in a bull versus a bear market. A standard sublinear-regret algorithm, competing…

9 thms4 active users
🏆Completed
Convex OptimizationProbabilityRandom Matrix Theory+1·Captain: mikedeng1

High-Dimensional Probability III: Grothendieck's InequalityTextbook

Many hard combinatorial optimization problems — finding the maximum cut of a graph, deciding the ground state of an Ising spin system, bounding the correlation of a physical system — can be written as maximizing a bilinear form over sign vectors xi∈{−1,1}x_i \in \{-1, 1\}xi​∈{−1,1}. Exhaustive search over 2n2^n2n sign patterns is intractable, so practitioners relax the problem:…

13 thms4 active users
🏆Completed
ProbabilityRandom Matrix TheoryStatistics·Captain: mikedeng1

High-Dimensional Probability V: The Johnson-Lindenstrauss LemmaTextbook

Any dataset of NNN points can be described exactly by embedding it in Rn\mathbb R^nRn for nnn large enough — but a large nnn is expensive: nearest-neighbor search, clustering, and streaming algorithms all scale with the ambient dimension, not with NNN. The question that opens this mission is whether the dimension can be cut down while leaving the data's geometry — the…

7 thms4 active usersReviewed
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization III: Online Gradient DescentTextbook

Online convex optimization (OCO) models a repeated decision process: at each round a learner picks a point in a convex set, an adversary (or the world) reveals a convex cost function, the learner pays that cost at its own point, and the process repeats. No statistical assumption on the sequence of costs is made. This model, introduced by Zinkevich…

8 thms4 active users
🏆Completed
Convex OptimizationOptimization·Captain: mikedeng1

Introduction to Online Convex Optimization I: Learning from Expert Advice and the Hedge AlgorithmTextbook

Consider a decision maker who must choose, at each of TTT rounds, between two actions on the advice of NNN "experts," none of which is known in advance to be reliable. This is the prediction-from-expert-advice problem, introduced by Littlestone and Warmuth [Littlestone & Warmuth, The Weighted Majority Algorithm, FOCS 1989/Inf. Comput. 1994] and generalized to…

9 thms4 active usersReviewed
🏆Completed
Control TheoryFunctional Analysis·Captain: olivier

Universal Reservoir Computers from Non-Homogeneous State-Affine SystemsResearch Paper

A reservoir computer learns a dynamical input/output relation with a recurrent network whose internal weights are fixed once and never trained; only a linear readout on the state is fitted. The method works in practice — it is a standard tool for learning chaotic dynamics — but its justification requires an approximation theorem: the family of reservoirs…

8 thms4 active usersReviewed
🏆Completed
Operations ResearchOptimizationStatistics·Captain: Shuze Chen

Matrix Completion has No Spurious Local MinimumResearch Paper

Matrix completion — recovering a low-rank matrix M=ZZ⊤M = ZZ^\topM=ZZ⊤ from a small random subset of its entries — powers recommender systems and collaborative filtering. In practice it is solved by running (stochastic) gradient descent on the non-convex objective

14 thms4 active users
🏆Completed
Bandit AlgorithmsOperations Research·Captain: Shuze Chen

Bandit Algorithms XV: Partial MonitoringTextbook

Bandit feedback is only one point on a spectrum: a learner might see more than its own loss (full information) or less (a spam filter never learns what happened to mail it deleted). Chapter 37 of Lattimore–Szepesvári studies finite adversarial games G=(L,Φ)G = (\mathcal{L}, \Phi)G=(L,Φ) where the loss matrix and the feedback matrix are decoupled. The goal theorem is the…

16 thms4 active users
🏆Completed
Bandit AlgorithmsOperations Research·Captain: Shuze Chen

Bandit Algorithms X: Stochastic Linear Bandits and LinUCBTextbook

When actions are feature vectors and the mean reward is linear — Xt=⟨θ∗,At⟩+ηtX_t = \langle \theta_*, A_t\rangle + \eta_tXt​=⟨θ∗​,At​⟩+ηt​ — a bandit can generalize across arms: pulling one arm reveals information about all of them. Chapter 19 of Lattimore–Szepesvári carries the optimism principle into this setting: LinUCB (a.k.a. OFUL) plays the action maximizing …

8 thms4 active usersReviewed
🏆Completed
Bandit AlgorithmsOperations Research·Captain: Shuze Chen

Bandit Algorithms IX: Self-Normalized Concentration and Optimal DesignTextbook

Least-squares estimation from adaptively collected data is the statistical heart of linear bandits: the actions AtA_tAt​ depend on past noise, so classical fixed-design theory does not apply. Chapter 20 of Lattimore–Szepesvári resolves this with the method of mixtures: the process Mt(x)=exp⁡(⟨x,St⟩−12∥x∥Vt(λ)2)M_t(x) = \exp(\langle x, S_t\rangle - \frac{1}{2}\|x\|^2_{V_t(\lambda)})Mt​(x)=exp(⟨x,St​⟩−21​∥x∥Vt​(λ)2​) is a…

9 thms4 active usersReviewed
Dynamical SystemsProbability·Captain: iris

Mitigating error propagation in chaotic dynamics via inference-time noising: a view from backward samplingResearch Paper

Machine-learned models of dynamical systems are usually trained to predict the next state and are then applied repeatedly to generate whole trajectories. For chaotic systems this is fragile: small one-step errors are amplified by the dynamics, and the states visited by the rollout drift away from the training distribution. Yet individual trajectories…

40 thms3 active usersReviewed
Computational ComplexityProbabilityTheoretical Computer Science·Captain: iris

Genetic Algorithms with Optimization Guided OperatorsResearch Paper

Systems such as FunSearch and AlphaEvolve search for programs and constructions by repeatedly asking a language model for new candidates, for local revisions of a stored candidate, and for combinations of several stored candidates. A. Brandenberger, I. Doron-Arad and E. Mossel (Genetic Algorithms with Optimization Guided Operators, STOC 2027 submission; an…

26 thms3 active usersReviewed
ProbabilityStatisticsTheoretical Computer Science·Captain: Shuze Chen

Posterior Sampling in Gaussian Mixture Models Beyond the Clusterability ThresholdResearch Paper

Bayesian mixture models are the standard probabilistic description of clustering: each data point carries a hidden label saying which component generated it, and the scientific object is the posterior over labels given the data. In practice the posterior is explored with Gibbs samplers and related Markov chains, which are known to get trapped between label…

21 thms3 active usersReviewed
Graph TheoryProbabilityStatistics·Captain: Shuze Chen

Supervised Learning with Hidden CorrespondencesResearch Paper

In many data sets the correspondence between inputs and labels is lost or only partly known: records are linked across databases without a shared identifier, labels are released only in aggregate over bags of examples, or a privacy mechanism shuffles them. The learner then sees inputs X1,…,XnX_1,\dots,X_nX1​,…,Xn​ and labels Y1,…,YnY_1,\dots,Y_nY1​,…,Yn​ together with a graph of plausible…

24 thms3 active usersReviewed
🏆Completed
CombinatoricsProbabilityTheoretical Computer Science·Captain: Shuze Chen

Algorithmic Principles For Multiclass Learning Are Hard To Come By: Limits of Regularization and Proper LearningResearch Paper

In binary classification, every PAC learnable class can be learned by a proper learner, one that always outputs a member of the class: empirical risk minimization does it, and learnability is characterized by the VC dimension. With many labels the picture changes. Daniely and Shalev-Shwartz introduced the DS dimension and conjectured that it characterizes…

11 thms3 active usersReviewed
Linear algebraProbabilityTheoretical Computer Science·Captain: Shuze Chen

Nearly Optimal Strong Coresets for ℓp\ell_pℓp​ Subspace ApproximationResearch Paper

Low-rank approximation under robust losses is a basic primitive of data analysis. Given nnn points in Rd\mathbb R^dRd, the ℓp subspace approximation problem asks for a subspace of dimension at most kkk minimizing the sum of ppp-th powers of the Euclidean distances of the points to it. For p=2p=2p=2 this is principal component analysis; for 1≤p<21\le p<21≤p<2 it is less sensitive to…

24 thms3 active usersReviewed
Complexity TheoryNumerical AnalysisTheoretical Computer Science·Captain: Shuze Chen

The Cancellation-Free Cost of Hessian Diagonals: Exact Counts, Formulas and InterfacesResearch Paper

Automatic differentiation computes the gradient of a function given as a program at a constant factor over one evaluation [Baur and Strassen 1983; Griewank and Walther 2008]. Second-order quantities are harder. The Hessian diagonal (∂2f/∂xi2)i(\partial^2 f/\partial x_i^2)_i(∂2f/∂xi2​)i​ and its sum, the Laplacian, appear in physics-informed neural networks, variational Monte Carlo…

15 thms3 active usersReviewed
PreviousPage 11 of 36Next

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me