arXiv subjects

Aditya Karan

Publications and source records attributed to Aditya Karan.

4 recordsLinked to original sources

arXiv2025

Sync or Sink: Bounds on Algorithmic Collective Action with Noise and Multiple Groups

Collective action against algorithmic systems provides an opportunity for a small group of individuals to strategically manipulate their data to get specific outcomes, from classification to recommendation models. This effectiveness will invite more growth of this type of coordinated actions, both in the size and the number of distinct collectives. With a small group, however, coordination is key. Currently, there is no formal analysis of how coordination challenges within a collective can impact downstream outcomes, or how multiple collectives may affect each other's success. In this work, we aim to provide guarantees on the success of collective action in the presence of both coordination noise and multiple groups. Our insight is that data generated by either multiple collectives or by coordination noise can be viewed as originating from multiple data distributions. Using this framing, we derive bounds on the success of collective action. We conduct experiments to study the effects of noise on collective action. We find that sufficiently high levels of noise can reduce the success of collective action. In certain scenarios, large noise can sink a collective success rate from $100\%$ to just under $60\%$. We identify potential trade-offs between collective size and coordination noise; for example, a collective that is twice as big but with four times more noise experiencing worse outcomes than the smaller, more coordinated one. This work highlights the importance of understanding nuanced dynamics of strategic behavior in algorithmic systems.

physics.soc-ph↗

arXiv2025

Algorithmic Collective Action with Two Collectives

Given that data-dependent algorithmic systems have become impactful in more domains of life, the need for individuals to promote their own interests and hold algorithms accountable has grown. To have meaningful influence, individuals must band together to engage in collective action. Groups that engage in such algorithmic collective action are likely to vary in size, membership characteristics, and crucially, objectives. In this work, we introduce a first of a kind framework for studying collective action with two or more collectives that strategically behave to manipulate data-driven systems. With more than one collective acting on a system, unexpected interactions may occur. We use this framework to conduct experiments with language model-based classifiers and recommender systems where two collectives each attempt to achieve their own individual objectives. We examine how differing objectives, strategies, sizes, and homogeneity can impact a collective's efficacy. We find that the unintentional interactions between collectives can be quite significant; a collective acting in isolation may be able to achieve their objective (e.g., improve classification outcomes for themselves or promote a particular item), but when a second collective acts simultaneously, the efficacy of the first group drops by as much as $75\%$. We find that, in the recommender system context, neither fully heterogeneous nor fully homogeneous collectives stand out as most efficacious and that heterogeneity's impact is secondary compared to collective size. Our results signal the need for more transparency in both the underlying algorithmic models and the different behaviors individuals or collectives may take on these systems. This approach also allows collectives to hold algorithmic system developers accountable and provides a framework for people to actively use their own data to promote their own interests.

cs.CY↗

arXiv2024

Parks: A Doubly Infinite Family of NP-Complete Puzzles and Generalizations of A002464

The Parks Puzzle is a paper-and-pencil puzzle game that is classically played on a square grid with different colored regions (the parks). The player needs to place a certain number of "trees" in each row, column, and park such that none are adjacent, even diagonally. We define a doubly-infinite family of such puzzles, the $(c, r)$-tree Parks puzzles, where there need be $c$ trees per column and $r$ per row. We then prove that for each $c$ and $r$ the set of $(c, r)$-tree puzzles is NP-complete. For each $c$ and $r$, there is a sequence of possible board sizes $m \times n$, and the number of possible puzzle solutions for these board sizes is a doubly-infinite generalization of OEIS sequence A002464, which itself describes the case $c = r = 1$. This connects the Parks puzzle to chess-based puzzle problems, as the sequence describes the number of ways to place non-attacking kings on a chessboard so that there is exactly one in each column and row (i.e. to place non-attacking dragon kings in shogi). These findings add yet another puzzle to the set of chess puzzles and expands the list of known NP-complete problems described.

cs.CC↗

arXiv2024

Oh the Prices You'll See: Designing a Fair Exchange System to Mitigate Personalized Pricing

Many online marketplaces personalize prices based on consumer attributes. Since these prices are private, consumers may be unaware that they have spent more on a good than the lowest possible price, and cannot easily take action to pay less. In this paper, we introduce a fairness-centered exchange system that takes advantage of personalized pricing, while still allowing consumers to individually benefit. Our system produces a matching of consumers to promote trading; the lower-paying consumer buys the good for the higher-paying consumer for some fee. We explore various modeling choices and fairness targets to determine which schema will leave consumers best off, while also earning revenue for the system itself. We show that when consumers individually negotiate the transaction price, and our fairness objective is to minimize mean net cost, we are able to achieve the most fair outcomes. Conversely, when transaction prices are centrally set, consumers are often unwilling to transact. When price dispersion (or range) is high, the system can reduce the mean net cost to each individual by $66\%$, or the mean net cost to a group by $69\%$. We find that a high dispersion of original prices is necessary for our system to be viable. Higher dispersion can actually lead to decreased net price paid by consumers, and act as a check against extreme personalization, increasing seller accountability. Our results provide theoretical evidence that such a system could improve fairness for consumers while sustaining itself financially.

cs.GT↗