Searcharxiv⌕ Search

arXiv subjects

Jiatian Wu

Publications and source records attributed to Jiatian Wu.

3 recordsLinked to original sources

Serving-Aligned Entire-Space Conversion Modeling via Click-Through Rate Knowledge Distillation

Recommendation systems may use predicted click-through conversion rate (pCTCVR), the product of predicted click-through rate (pCTR) and post-click conversion rate (pCVR), to guide ranking decisions. Many pCVR models build on the Entire Space Multi-Task Model (ESMM), which jointly trains pCVR and auxiliary pCTR towers using pCTR and pCTCVR losses, but deploys only the pCVR tower. At serving time, this tower may be paired with a separately trained pCTR model, creating a training-serving mismatch in pCTCVR when its predictions differ from those of the auxiliary pCTR tower. In this paper, we use knowledge distillation (KD) from logged serving pCTR scores to supervise ESMM's auxiliary pCTR tower and improve training-serving pCTCVR alignment. We denote variants by KD-$α$, where $α$ controls the hard-label loss weight. On the Ali-CCP dataset, KD-0.3, the best variant on serving pCTCVR, increases ROC-AUC not only for serving pCTCVR but also for pCVR by 2.1% and 1.6%, respectively, and PR-AUC by 9.9% and 8.8%, respectively, relative to the ESMM baseline. We also test negative downsampling with 5% nonclick retention. Downsampled KD-0.3 increases mean AUC for serving pCTCVR and pCVR by 1.2% and 0.9%, respectively, -and PR-AUC by 8.5% and 4.5%, respectively, relative to downsampled ESMM.

cs.LG↗

Combining Quasiparticle Self-Consistent $GW$ and Machine-Learned DFT+$U$ to Assess Half-Metallicity in Co- and Ni-Based Heuslers

Half-metallic Heusler compounds are of significant interest for spintronics. For device fabrication, compounds that can be epitaxially grown on III-V semiconductors are particularly attractive. We present a first-principles investigation of four Co-based and two Ni-based Heusler compounds that are lattice-matched to InAs. The results of density functional theory (DFT) using semi-local and hybrid functionals are compared to quasiparticle self-consistent $GW$ (QP$GW$). We also consider DFT with machine-learned Hubbard $U$ corrections [npj Computational Materials 6, 180 (2020)] with a new Bayesian optimization (BO) objective function to determine the $U$ values that yield the closest agreement with the QP$GW$ band structure and magnetic moments. We find that DFT+$U$(BO) can adequately reproduce the key QP$GW$ features in most cases. Our results reveal a strong method dependence of the degree of spin polarization at the Fermi level and, in some cases, even the dominant spin channel (majority or minority). Of the materials studied here, Co$_2$TiSn and Co$_2$ZrAl are the most likely to be half-metals.

cond-mat.mtrl-sci↗

Optical control of valley Zeeman effect through many-exciton interactions

Charge carriers in two-dimensional transition metal dichalcogenides (TMDs), such as WSe$_2$, have their spin and valley-pseudospin locked into an optically-addressable index that is proposed as a basis for future information processing. The manipulation of this spin-valley index requires tuning its energy, typically through external magnetic field (B), which is cumbersome. Thus, other efficient routes like all-optical control of spin-valley index are desirable. Here, we show that many-body interactions amongst interlayer excitons in WSe$_2$/MoSe$_2$ heterobilayer induce a steady-state valley Zeeman splitting corresponding to B $\sim$ 6 Tesla. This anomalous splitting, present at incident powers as low as $μ$Ws, increases with power and enhances, suppresses or even flips the sign of a B-induced splitting. Moreover, the $\it{g}$-factor of valley Zeeman splitting can be tuned by $\sim$ 30 $\%$ with incident power. In addition to valleytronics, our results are relevant for achieving optical non-reciprocity using two-dimensional materials.

cond-mat.mes-hall↗