Searcharxiv⌕ Search

arXiv subjects

Jingjing Duan

Publications and source records attributed to Jingjing Duan.

4 recordsLinked to original sources

Quadratic Elements in Finite $p$-Groups and Applications to Oliver's $p$-Group Conjecture

We establish a property of quadratic elements in faithful modules for finite $p$-groups. Let $p$ be an odd prime, and suppose that $a$ is a quadratic element of an abelian subgroup $L$ normalized by $b$. Writing $a_0=a$ and $a_i=[a_{i-1},b]$, we prove that if, for some $r$ with $2\leq r\leq(p+1)/2$, one has $a_{r-1}\neq1$ and $a_r=1$, then $a_{r-1}$ is quadratic. As an application, we prove that every finite $p$-group of nilpotency class at most $(p+5)/2$ satisfies the Quadratic Conjecture. Consequently, Oliver's $p$-group conjecture holds for a finite $p$-group $S$ whenever $S/\mathfrak X(S)$ has nilpotency class at most $(p+5)/2$.

math.GR↗

TS-GEAR: Taobao Shangou GEnerative Ad Retrieval

On-demand delivery platforms such as DoorDash and Uber Eats impose a constraint absent in traditional e-commerce: merchants beyond the user's delivery range are invalid candidates rather than merely low-ranked. Generative retrieval has recently shown strong potential in large-scale retrieval systems. Adapting it to this setting, however, poses three challenges. First, existing semantic IDs encode geography only as a soft signal, so reachability can only rely on a non-differentiable post-generation hard truncation. Second, user preferences vary substantially across both regions and queries, so aggregating historical behaviors without conditioning on geography and the current query yields biased interest estimates. Third, the retrieved candidate set is evaluated by its collective value, yet autoregressive training scores each sequence independently, with no set-level signal to discourage redundant near-duplicates. To address the challenges, we propose TS-GEAR, a unified generative retrieval framework for on-demand delivery: (1) Geo-Semantic ID splits merchant identifiers into geographic prefixes and semantic suffixes, making reachability a prediction target rather than a post-hoc filter; (2) Query-Geo Target Attention aggregates historical behaviors using the real-time query and location as dual targets, while encoding each behavior's own location into its attention key; (3) GMPO (Group Marginal Policy Optimization) rewrites advantages using a deduplicated group reward and marginal advantage within GRPO, avoiding auxiliary diversity objectives. Deployed online with a lightweight 0.1B backbone, TS-GEAR serves over 150 million queries per day in the search advertising pipeline of Taobao Shangou, consistently outperforming state-of-the-art generative retrieval methods on both HitRate and geographic validity, and achieving 3.8% gains in ad impressions and 4.2% in ad revenue.

cs.IR↗

The_Quadratic_Conjecture_and_p-groups_of_maximal_class

The Quadratic Conjecture for finite p-groups, which is equivalent to the Strong Oliver Conjecture, asserts that if a finite p-group G acts faithfully on an F-module V, then the center Z(G) contains a quadratic element. In this paper, we investigate the structure of maximal class p-groups and establish that they satisfy the Quadratic Conjecture.

math.GR↗

On quadratic conjecture

Quadratic conjecture is a strengthening of oliver's $p$-group conjecture. Let $G$ be a $p$-group of maximal class of order $p^n$. We prove that if $n\le 8$ or $n\ge \max\{2p-6,p+2\}$ then $G$ satisfies Quadratic Conjecture. Hence quadratic conjecture holds if $G$ is a $p$-group of maximal class where $p\le 7$.

math.GR↗