Industrial generative recommendation systems operating over discrete Semantic IDs (SIDs) are largely behavior-driven, and thus struggle to proactively activate latent demand before explicit user signals emerge, leading to intent cold-start. To address this, we propose RIGER (Reasoning-grounded Intent injection for GE nerative Recommendation), a deployable two-stage framework that integrates offline large language model (LLM) reasoning into an online generative recommender under strict latency constraints. Offline, to ensure scalable deployment, we distill the latent-intent inference capability of a strong LLM into a lightweight forecasting model using an automated data curation pipeline---leveraging judge-guided prompt calibration and future-query-guided rejection filtering. Online, to bridge the representation mismatch between free-form textual intents and the discrete SID token space, predicted intents are converted into SID-native tokens through a behavior-grounded mapping and injected into the deployed decoder-only retrieval backbone. We further fine-tune the model with beam-aware GRPO, introducing a hierarchical intent-alignment exploration reward in SID space while preserving exploitation behavior through KL regularization. Offline evaluations demonstrate a substantial increase in intent-aligned density and diversity with only a marginal reduction in hindsight recall, indicating that RIGER effectively enhances proactive intent exploration while preserving its capability to exploit historical behaviors. In a large-scale e-commerce display advertising system, RIGER improves clicks by 1.6% and advertiser spend by 1.3%.
Xusong Chen, Peini Guo, Fang Liu et al.· Annual International ACM SIG...· 0 citations
BACKGROUND
Colorectal cancer (CRC) is a common malignancy worldwide, and its progression is associated with dysregulated signaling pathways. Fibroblast growth factor receptor 2 (FGFR2) is frequently overexpressed in CRC and correlated with poor prognosis, but whether it mediates hypoxia-inducible factor-1α (HIF-1α) to promote tumor progression remains unclear.
METHODS
The expression and prognostic value of FGFR2 and HIF-1α in CRC were analyzed via bioinformatics. Clinicopathological specimens were obtained and immunohistochemical staining was performed to explore the correlations between FGFR2 and HIF-1α expression and the clinicopathological features of CRC patients. Using CRC cell lines (HCT116 and NCI-H716), cell proliferation, transwell, wound healing, western blotting, gene activation and knockdown experiments were used to assess the effects of FGFR2 expression on cell proliferation, migration and invasion, as well as phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) pathway activity and HIF-1α expression level. A mouse xenotransplantation model was used to further clarify the regulatory effects of changes in FGFR2 expression on PI3K/AKT/mTOR pathway activity, HIF-1α expression level and CRC tumor growth in vivo.
RESULTS
In this study, immunohistochemical staining of tumor tissues from 72 patients with CRC confirmed that the expression level of HIF-1α in cancer tissues was higher than that in normal tissues. Bioinformatics analysis and clinicopathological analysis confirmed that the overexpression of the FGFR2 and HIF-1α genes were the independent risk factors affecting the prognosis of CRC and that the overexpression of HIF-1α was positively correlated with the increase of tumor size and FGFR2 expression. The in vitro results of this study show that FGFR2 mediates HIF-1α and affects CRC cell proliferation, migration and invasion, which is consistent with the results of bioinformatics analysis. In vitro experiments involving FGFR2 overactivation and knockdown demonstrated that FGFR2 primarily functions in regulating the PI3K/AKT/mTOR signaling pathway, which was further verified by xenotransplantation mouse model. Knockdown of FGFR2 expression curbed the growth of CRC tumors through downregulating the PI3K/AKT/mTOR signaling pathway and HIF-1α expression level.
CONCLUSION
FGFR2 mediates the expression of HIF-1α via activating the PI3K/AKT/mTOR signaling pathway, thereby enhancing the proliferation, migration, and invasion of CRC cells. These findings suggest that FGFR2 may serve as potential therapeutic targets for the treatment of CRC.
Yi-Yang Hu, Liu-Xu Yao, Yuhong Li et al.· Tissue & Cell· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
