-
Learning from data with structured missingness
Authors:
Robin Mitra,
Sarah F. McGough,
Tapabrata Chakraborti,
Chris Holmes,
Ryan Copping,
Niels Hagenbuch,
Stefanie Biedermann,
Jack Noonan,
Brieuc Lehmann,
Aditi Shenvi,
Xuan Vinh Doan,
David Leslie,
Ginestra Bianconi,
Ruben Sanchez-Garcia,
Alisha Davies,
Maxine Mackintosh,
Eleni-Rosalina Andrinopoulou,
Anahid Basiri,
Chris Harbron,
Ben D. MacArthur
Abstract:
Missing data are an unavoidable complication in many machine learning tasks. When data are `missing at random' there exist a range of tools and techniques to deal with the issue. However, as machine learning studies become more ambitious, and seek to learn from ever-larger volumes of heterogeneous data, an increasingly encountered problem arises in which missing values exhibit an association or st…
▽ More
Missing data are an unavoidable complication in many machine learning tasks. When data are `missing at random' there exist a range of tools and techniques to deal with the issue. However, as machine learning studies become more ambitious, and seek to learn from ever-larger volumes of heterogeneous data, an increasingly encountered problem arises in which missing values exhibit an association or structure, either explicitly or implicitly. Such `structured missingness' raises a range of challenges that have not yet been systematically addressed, and presents a fundamental hindrance to machine learning at scale. Here, we outline the current literature and propose a set of grand challenges in learning from data with structured missingness.
△ Less
Submitted 3 April, 2023;
originally announced April 2023.
-
Fairness Criteria for Allocating Indivisible Chores: Connections and Efficiencies
Authors:
Ankang Sun,
Bo Chen,
Xuan Vinh Doan
Abstract:
We study several fairness notions in allocating indivisible chores (i.e., items with non-positive values) to agents who have additive and submodular cost functions. The fairness criteria we are concern with are envy-free up to any item (EFX), envy-free up to one item (EF1), maximin share (MMS), and pairwise maximin share (PMMS), which are proposed as relaxations of envy-freeness in the setting of…
▽ More
We study several fairness notions in allocating indivisible chores (i.e., items with non-positive values) to agents who have additive and submodular cost functions. The fairness criteria we are concern with are envy-free up to any item (EFX), envy-free up to one item (EF1), maximin share (MMS), and pairwise maximin share (PMMS), which are proposed as relaxations of envy-freeness in the setting of additive cost functions. For allocations under each fairness criterion, we establish their approximation guarantee for other fairness criteria. Under the additive setting, our results show strong connections between these fairness criteria and, at the same time, reveal intrinsic differences between goods allocation and chores allocation. However, such strong relationships cannot be inherited by the submodular setting, under which PMMS and MMS are no longer relaxations of envy-freeness and, even worse, few non-trivial guarantees exist. We also investigate efficiency loss under these fairness constraints and establish their prices of fairness.
△ Less
Submitted 28 September, 2021; v1 submitted 18 January, 2021;
originally announced January 2021.
-
Robust Newsvendor Games with Ambiguity in Demand Distributions
Authors:
Xuan Vinh Doan,
Tri-Dung Nguyen
Abstract:
We investigate newsvendor games whose payoff function is uncertain due to ambiguity in demand distributions. We discuss the concept of stability under uncertainty and introduce solution concepts for robust cooperative games which could be applied to these newsvendor games. Properties and numerical schemes for finding core solutions of robust newsvendor games are presented.
We investigate newsvendor games whose payoff function is uncertain due to ambiguity in demand distributions. We discuss the concept of stability under uncertainty and introduce solution concepts for robust cooperative games which could be applied to these newsvendor games. Properties and numerical schemes for finding core solutions of robust newsvendor games are presented.
△ Less
Submitted 28 January, 2016; v1 submitted 24 March, 2014;
originally announced March 2014.
-
Price of Anarchy for Non-atomic Congestion Games with Stochastic Demands
Authors:
Chenlan Wang,
Xuan Vinh Doan,
Bo Chen
Abstract:
We generalize the notions of user equilibrium and system optimum to non-atomic congestion games with stochastic demands. We establish upper bounds on the price of anarchy for three different settings of link cost functions and demand distributions, namely, (a) affine cost functions and general distributions, (b) polynomial cost functions and general positive-valued distributions, and (c) polynomia…
▽ More
We generalize the notions of user equilibrium and system optimum to non-atomic congestion games with stochastic demands. We establish upper bounds on the price of anarchy for three different settings of link cost functions and demand distributions, namely, (a) affine cost functions and general distributions, (b) polynomial cost functions and general positive-valued distributions, and (c) polynomial cost functions and the normal distributions. All the upper bounds are tight in some special cases, including the case of deterministic demands.
△ Less
Submitted 17 October, 2013;
originally announced October 2013.
