DUET: Dual User Embedding Transformers for Offsite Conversion Prediction
Researchers have proposed DUET (Dual User Embedding Transformers), a machine learning framework that uses two separate transformer encoders to predict whether users will complete purchases or other conversions on external sites after clicking ads. The system addresses a core challenge in recommendation systems: click data is plentiful and immediate, while conversion data is sparse, delayed, and often unattributed. The approach reportedly achieves up to 0.38% normalized entropy reduction over the strongest baseline and showed consistent gains in live A/B testing.
A team of researchers, primarily affiliated with an industry setting given the large author list, has introduced DUET (Dual User Embedding Transformers), a pre-training framework designed to improve offsite conversion rate (OCVR) prediction in computational recommendation systems. The central innovation is explicitly splitting user behavioral data into two streams — clicks and conversions — and training dedicated transformer encoders optimized for each stream's statistical properties: multi-layer self-attention for the dense click stream and interleaved cross- and self-attention for the sparse conversion stream. Prior approaches applied a single, undifferentiated encoder to both data types, which the authors argue fails to account for the fundamental statistical disparities between the two signals. The complementary embeddings produced by the two encoders are then jointly fed into a downstream ranking model, with the system designed to remain within real-world serving-latency constraints. Evaluation results show up to 0.38% normalized entropy (NE) reduction relative to the strongest baseline, and live A/B testing demonstrated consistent improvements in OCVR prediction accuracy. The paper was submitted to arXiv on June 8, 2026, and has not yet undergone formal peer review.
What's missing
The paper has not yet been peer-reviewed, as it is a preprint. Key open questions include: the magnitude and statistical significance of A/B test improvements beyond the NE metric, computational costs of training and serving two encoders versus one, and how the system handles cold-start scenarios for new users with little click or conversion history.
What different sources said
- arXiv cs.LGCenter
DUET -- Dual User Embedding Transformers for Offsite Conversion Prediction
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.