arXiv subjects

He Guo

Publications and source records attributed to He Guo.

At least 19 recordsLinked to original sources

arXiv2026

TMallGS: Scaling Unified Feature and Sequence Modeling for Generative E-commerce Search

In industrial search and ranking systems, Click-Through Rate (CTR) prediction is shifting from traditional Deep Learning Recommendation Models (DLRM) toward unified, compute-intensive Transformer architectures. This transition is driven by the need to improve Model FLOPs Utilization (MFU) and achieve predictable gains through scaling laws. However, existing approaches such as OneTrans and Climber often adopt an all-in-tokenization strategy when adapting Large Language Model (LLM) architectures, overlooking the heterogeneous nature of ranking features. We propose TmallGS, a scalable ranking architecture for Tmall search. TmallGS includes five key components: (1) Hierarchical Distribution-Calibrated Tokenization, which combines Field-wise Saliency Reweighting (FSR) and Distribution-Calibrated Projection (DCP) to map diverse features into optimized subspaces; (2) a Field-Adaptive Gated Transformer Backbone with per-field QKV projections and noise-adaptive gating for refined semantic interaction; (3) Decoupled FiLM Late Fusion to preserve explicit high-frequency signals; (4) a Context-Aware Bias Net to decouple systemic bias from user intent; and (5) Error-Aware Progressive Training with dynamically weighted losses for robust learning. Extensive offline experiments and online A/B tests on Tmall Search show that TmallGS improves training throughput and achieves substantial gains in UCTCVR and GMV.

cs.IR↗

arXiv2026

Learning to Forget: Satiation-Aware Long-Sequence Transducers for Mitigating Post-Purchase Redundancy

Sequential recommendation models predominantly interpret user interactions as positive signals for preference accumulation. However, in e-commerce scenarios, a purchase action often signifies the termination of a specific intent ("Interest Exit") rather than its continuation. Existing models overlook this distinction, suffering from Action-Intent Asymmetry, which leads to severe post-purchase redundancy. In this paper, we propose the Satiation-Aware Mechanism (SAM), an end-to-end framework designed to explicitly model the lifecycle of user interests. SAM incorporates three key components: (1) A Dual-path Cross-Attention architecture that retroactively suppresses historical clicks associated with a fulfilled intent while simultaneously retrieving personalized replenishment rhythms from long-term purchase history; (2) An Adaptive Satiation Gating Unit (ASGU) that generates a time-sensitive soft mask to inhibit satisfied interests immediately after purchase and gradually "re-awaken" them as the predicted repurchase cycle approaches; and (3) A self-supervised Time-to-Next-Purchase (TTNP) auxiliary task to learn latent product lifecycles without manual annotation. Extensive offline experiments on industrial datasets and online A/B testing demonstrate that SAM significantly reduces the Post-Purchase Repeat Rate (PPRR) by over 60%.

cs.IR↗

arXiv2026

From Head to Tail: Asymmetric Knowledge Transfer in Long-tail Recommendation with Generative Semantic IDs

Long-tail recommendation in real-world e-commerce platforms remains challenging due to severe data imbalance. Existing methods often struggle to combine content-based multimodal features with collaborative signals. Many of these methods also ignore an important asymmetry in knowledge transfer between head and tail IDs: noisy signals from tail IDs can hurt representation learning for head IDs. This paper presents AKT-Rec, a framework for Asymmetric Knowledge Transfer in long-tail Recommendation that uses LLM-generated semantic IDs. AKT-Rec uses Multimodal LLMs (MLLMs) with supervised fine-tuning to align content representations with collaborative information for both items and users, producing semantic representations. It then discretizes these representations into semantic IDs with a Residual-Quantized VAE (RQ-VAE), which yields semantic clusters of similar entities. AKT-Rec has two main components: (1) Cluster-Guided Adaptive Embedding, which decomposes each ID representation into a cluster-level embedding that captures shared semantics and an individual embedding. Through an asymmetric contrastive objective and an activity-aware gating mechanism, this module directs knowledge transfer from head to tail IDs. (2) Hierarchical Feature Aggregation, which builds parallel feature views and adaptively fuses them to optimize predictions for samples with varying activity levels. Extensive experiments on a large-scale industrial dataset and online A/B testing on the Alibaba Tmall platform demonstrate the effectiveness of AKT-Rec. AKT-Rec improves offline performance by 0.35% in AUC and 1.53% in GAUC, outperforming several competitive baselines. In online A/B testing, AKT-Rec achieves a 2.76% increase in CTR and a 3.47% increase in GMV, validating its utility in real-world production environments.

cs.IR↗

arXiv2026

Random 0/1-polytopes expand rapidly

A 0/1-polytope is the convex hull of a subset $V\subseteq \{0,1\}^n$. A celebrated conjecture of Mihail and Vazirani asserts that the graph of every 0/1-polytope has edge-expansion at least 1. In this paper, we show that typical 0/1-polytopes have significantly stronger expansion. Specifically, if $V$ is formed by sampling each vertex of $\{0,1\}^n$ independently with constant probability $p$, then with high probability the edge-expansion is $\Theta(n)$ for $p \in (1/2, 1)$, and $n^{\Theta(\log \log n)}$ for $p \in (0, 1/2)$. This improves the previously best known bound $\Omega(1)$ due to Ferber, Krivelevich, Sales and Samotij.

math.CO↗

arXiv2025

Short rainbow cycles for families of small edge sets

In 2019, Aharoni proposed a conjecture generalizing the Caceetta-H\"aggkvist conjecture: if an $n$-vertex graph $G$ admits an edge coloring (not necessarily proper) with $n$ colors such that each color class has size at least $r$, then $G$ contains a rainbow cycle of length at most $\lceil n/r\rceil$. Recent works \cite{AG2023,ABCGZ2023,G2025} have shown that if a constant fraction of the color classes are non-star, then the rainbow girth is $O(\log n)$. In this note, we extend these results, and we show that even a small fraction of non-star color classes suffices to ensure logarithmic rainbow girth. We also prove that the logarithmic bound is of the right order of magnitude. Moreover, we determine the threshold fraction between the types of color classes at which the rainbow girth transitions from linear to logarithmic.

math.CO↗

arXiv2024

big.LITTLE Vision Transformer for Efficient Visual Recognition

In this paper, we introduce the big.LITTLE Vision Transformer, an innovative architecture aimed at achieving efficient visual recognition. This dual-transformer system is composed of two distinct blocks: the big performance block, characterized by its high capacity and substantial computational demands, and the LITTLE efficiency block, designed for speed with lower capacity. The key innovation of our approach lies in its dynamic inference mechanism. When processing an image, our system determines the importance of each token and allocates them accordingly: essential tokens are processed by the high-performance big model, while less critical tokens are handled by the more efficient little model. This selective processing significantly reduces computational load without sacrificing the overall performance of the model, as it ensures that detailed analysis is reserved for the most important information. To validate the effectiveness of our big.LITTLE Vision Transformer, we conducted comprehensive experiments on image classification and segment anything task. Our results demonstrate that the big.LITTLE architecture not only maintains high accuracy but also achieves substantial computational savings. Specifically, our approach enables the efficient handling of large-scale visual recognition tasks by dynamically balancing the trade-offs between performance and efficiency. The success of our method underscores the potential of hybrid models in optimizing both computation and performance in visual recognition tasks, paving the way for more practical and scalable deployment of advanced neural networks in real-world applications.

cs.CV↗

arXiv2024

Coloring the intersection of two matroids

A result [The intersection of a matroid and a simplicial complex, Trans. Amer. Math. Soc. 358] from 2006 of Aharoni and the first author of this paper states that for any two positive integers $p,q$, where $p$ divides $q$, if a matroid $\mathcal{M}$ is $p$-colorable and a matroid $\mathcal{N}$ is $q$-colorable then $\mathcal{M} \cap \mathcal{N}$ is $(p+q)$-colorable. In this paper we show that the assumption that $p$ divides $q$ is in fact redundant, and we also prove that $\mathcal{M} \cap \mathcal{N}$ is even $p+q$ list-colorable. The result uses topology and relies on a new parameter yielding a lower bound for the topological connectivity of the intersection of two matroids.

math.CO↗

arXiv2024

Coloring, list coloring, and fractional coloring in intersections of matroids

It is known that in matroids the difference between the chromatic number and the fractional chromatic number is smaller than 1, and that the list chromatic number is equal to the chromatic number. We investigate the gap within these pairs of parameters for hypergraphs that are the intersection of a given number k of matroids. We prove that in such hypergraphs the list chromatic number is at most k times the chromatic number and at most 2k-1 times the maximum chromatic number among the k matroids. We study the relationship between three polytopes associated with k-sets of matroids, and connect them to bounds on the fractional chromatic number of the intersection of the members of the k-set. This also connects to bounds on the matroidal matching and covering number of the intersection of the members of the k-set. The tools used are in part topological.

math.CO↗

arXiv2024

The list chromatic number of the intersection of two generalized partition matroids

A famous theorem of Galvin states that the list chromatic number of the intersection of two partition matroids equals its chromatic number. Kiraly and Berczi et. al. conjectured that this equality holds for any two matroids. We prove this conjecture and a conjecture by Aharoni--Berger for any two generalized partition matroids.

math.CO↗

arXiv2024

In-Context Matting

We introduce in-context matting, a novel task setting of image matting. Given a reference image of a certain foreground and guided priors such as points, scribbles, and masks, in-context matting enables automatic alpha estimation on a batch of target images of the same foreground category, without additional auxiliary input. This setting marries good performance in auxiliary input-based matting and ease of use in automatic matting, which finds a good trade-off between customization and automation. To overcome the key challenge of accurate foreground matching, we introduce IconMatting, an in-context matting model built upon a pre-trained text-to-image diffusion model. Conditioned on inter- and intra-similarity matching, IconMatting can make full use of reference context to generate accurate target alpha mattes. To benchmark the task, we also introduce a novel testing dataset ICM-$57$, covering 57 groups of real-world images. Quantitative and qualitative results on the ICM-57 testing set show that IconMatting rivals the accuracy of trimap-based matting while retaining the automation level akin to automatic matting. Code is available at https://github.com/tiny-smart/in-context-matting

cs.CV↗

arXiv2023

2-covers of wide Young diagrams

A Young diagram $Y$ is called wide if every sub-diagram $Z$ formed by a subset of the rows of $Y$ dominates $Z'$, the conjugate of $Z$. A Young diagram $Y$ is called Latin if its squares can be assigned numbers so that for each $i$, the $i$th row is filled injectively with the numbers $1, \ldots ,a_i$, where $a_i$ is the length of $i$th row of $Y$, and every column is also filled injectively. A conjecture of Chow and Taylor, publicized by Chow, Fan, Goemans, and Vondrak is that a wide Young diagram is Latin. We prove a dual version of the conjecture.

math.CO↗

arXiv2023

Looms

A pair $(A,B)$ of hypergraphs is called orthogonal if $|a \cap b|=1$ for every pair of edges $a \in A$ and $b \in B$. An orthogonal pair of hypergraphs is called a loom if each of its two members is the set of minimum covers of the other. Looms appear naturally in the context of a conjecture of Gy\'arf\'as and Lehel on the covering number of cross-intersecting hypergraphs. We study their properties and ways of construction, and prove special cases of a conjecture that if true would imply the Gy\'arf\'as--Lehel conjecture.

math.CO↗

arXiv2023

Tight infinite matrices

We give a simple proof of a recent result of Gollin and Jo\'o: if a possibly infinite system of homogeneous linear equations $A\vec{x} = \vec{0}$, where $A = (a_{i, j})$ is an $I \times J$ matrix, has only the trivial solution, then there exists an injection $\phi: J \to I$, such that $a_{\phi(j), j} \neq 0$ for all $j \in J$.

math.CO↗

arXiv2022

Short rainbow cycles for families of matchings and triangles

A generalization of the famous Caccetta--H\"aggkvist conjecture, suggested by Aharoni [Rainbow triangles and the Caccetta-H\"aggkvist conjecture, J. Graph Theory (2019)], is that any family $\mathcal{F}=(F_1, \ldots,F_n)$ of sets of edges in $K_n$, each of size $k$, has a rainbow cycle of length at most $\lceil \frac{n}{k}\rceil$. In [Rainbow cycles for families of matchings, Israel J. Math. (2023)] and [Non-uniform degrees and rainbow versions of the Caccetta-H\"aggkvist conjecture, SIAM J. Discrete Math. (2023)] it was shown that asymptotically this can be improved to $O(\log n)$ if all sets are matchings of size 2, or all are triangles. We show that the same is true in the mixed case, i.e., if each $F_i$ is either a matching of size 2 or a triangle. We also study the case that each $F_i$ is a matching of size 2 or a single edge, or each $F_i$ is a triangle or a single edge, and in each of these cases we determine the threshold proportion between the types, beyond which the rainbow girth goes from linear to logarithmic.

math.CO↗

arXiv2022

Visual Encoding and Debiasing for CTR Prediction

Extracting expressive visual features is crucial for accurate Click-Through-Rate (CTR) prediction in visual search advertising systems. Current commercial systems use off-the-shelf visual encoders to facilitate fast online service. However, the extracted visual features are coarse-grained and/or biased. In this paper, we present a visual encoding framework for CTR prediction to overcome these problems. The framework is based on contrastive learning which pulls positive pairs closer and pushes negative pairs apart in the visual feature space. To obtain fine-grained visual features,we present contrastive learning supervised by click through data to fine-tune the visual encoder. To reduce sample selection bias, firstly we train the visual encoder offline by leveraging both unbiased self-supervision and click supervision signals. Secondly, we incorporate a debiasing network in the online CTR predictor to adjust the visual features by contrasting high impression items with selected items with lower impressions.We deploy the framework in the visual sponsor search system at Alibaba. Offline experiments on billion-scale datasets and online experiments demonstrate that the proposed framework can make accurate and unbiased predictions.

cs.IR↗

arXiv2021

Rainbow cycles for families of matchings

Given a graph $G$ and a coloring of its edges, a subgraph of $G$ is called rainbow if its edges have distinct colors. The rainbow girth of an edge coloring of G is the minimum length of a rainbow cycle in G. A generalization of the famous Caccetta-H\"aggkvist conjecture, proposed by the first author, is that if in an coloring of the edge set of an $n$-vertex graph by $n$ colors, in which each color class is of size $k$, the rainbow girth is at most $\lceil \frac{n}{k} \rceil$. In the known examples for sharpness of this conjecture the color classes are stars, suggesting that when the color classes are matchings, the result may be improved. We show that the rainbow girth of $n$ matchings of size at least 2 is $O(\log n)$.

math.CO↗

arXiv2021

Non-uniform degrees and rainbow versions of the Caccetta-H\"aggkvist conjecture

The Caccetta-H\"aggkvist conjecture (denoted below CHC) states that the directed girth (the smallest length of a directed cycle) $dgirth(D)$ of a directed graph $D$ on $n$ vertices is at most $\lceil \frac{n}{\delta^+(D)}\rceil$, where $\delta^+(D)$ is the minimum out-degree of~$D$. We consider a version involving all out-degrees, not merely the minimum one, and prove that if $D$ does not contain a sink, then $dgirth(D) \le 2 \sum_{v\in V(D)} \frac{1}{deg^+(v)+1}$. In the spirit of a generalization of the CHC to rainbow cycles in \cite{ADH2019}, this suggests the conjecture that given non-empty sets $F_1, \ldots,F_n$ of edges of $K_n$, there exists a rainbow cycle of length at most $2\sum_{1\le i \le n}\frac{1}{|F_i|+1}$. We prove a bit stronger result when $1\le |F_i|\le 2$, thereby strengthening a result of DeVos et. al \cite{DDFGGHMM2021}. We prove a logarithmic bound on the rainbow girth in the case that the sets $F_i$ are triangles.

math.CO↗

arXiv2021

On the power of random greedy algorithms

In this paper we solve two problems of Esperet, Kang and Thomasse as well as Li concerning (i) induced bipartite subgraphs in triangle-free graphs and (ii) van der Waerden numbers. Each time random greedy algorithms allow us to go beyond the Lovasz Local Lemma or alteration method used in previous work, illustrating the power of the algorithmic approach to the probabilistic method.

math.CO↗