arXiv subjects

Jason Huang

Publications and source records attributed to Jason Huang.

8 recordsLinked to original sources

arXiv2026

Algorithmic Contract Design at Scale: Adaptive Peer Comparison for Enterprise Pricing

In enterprise software, a contract commits the customer to a usage volume over a fixed term in exchange for discounted pricing. These contracts are individually negotiated across many dimensions -- size, duration, industry, product mix, usage history -- and without a data-driven reference point, discounts tend to be overly generous. Manual governance review enforces discipline but at days-scale per contract, with inconsistency across reviewers and no real-time feedback to sellers. We present \emph{Contract Scoring}, a peer-based grading system deployed on every contract at Databricks. The system identifies empirically similar historical contracts via adaptive nearest neighbors over ensemble trees, where shared leaf membership defines a data-driven similarity learned from the discount target. It returns a letter grade with per-product-line breakdown in seconds; the underlying peer set is available to the centralized review team for audit. Sellers treat the grade as a contract design ``exit criterion'', iteratively adjusting discount structures until the grade reflects their intended tradeoff. Deployment evidence shows measurable discount discipline across the scored portfolio, with a commercially significant impact on revenue.

stat.AP↗

arXiv2026

Bias-Aware Confidence Intervals for Synthetic Control via Placebo-in-Time Bootstrap

Synthetic control (SC) methods are among the most widely used tools for causal inference without randomization. The standard Gaussian confidence interval around the estimated effect is simple, fast, and reliably directional when the treatment signal is strong, so practitioners default to it for good reason. Most treated populations, however, are bottom-heavy in intensity, and for them the SC model's systematic bias rivals or exceeds the signal even under good pre-treatment fit. Because this bias shares sign across units it does not average out, and the Gaussian confidence interval shrinks past it and converges on a wrong center. The failure is not imprecision but misdirection: a positive effect estimated as negligible is a missed opportunity, while a negligible effect estimated as significantly positive leads to continued investment in an intervention that is not working. No existing confidence interval for the SC effect measures this bias. We propose a placebo-in-time bootstrap that estimates the bias distribution directly from the observed panel. For each treated unit the procedure backdates the treatment onset and refits the SC model at each placebo onset; the resulting placebo gaps are draws from the same bias distribution that contaminates the real estimate, and bootstrapping them yields a critical value calibrated at the zero null. Because the method resamples realized model error rather than a hypothesized effect, coverage is trajectory-agnostic: it holds at fixed width regardless of how the true effect evolves over time.

stat.ME↗

arXiv2024

Measuring Consumer Sensitivity to Audio Advertising: A Long-Run Field Experiment on Pandora Internet Radio

A randomized experiment with almost 35 million Pandora listeners enables us to measure the sensitivity of consumers to advertising, an important topic of study in the era of ad-supported digital content provision. The experiment randomized listeners into nine treatment groups, each of which received a different level of audio advertising interrupting their music listening, with the highest treatment group receiving more than twice as many ads as the lowest treatment group. By maintaining consistent treatment assignment for 21 months, we measure long-run demand effects and find ad-load sensitivity three times greater than what we would have obtained from a month-long experiment. We show the negative impact on the number of hours listened, days listened, and probability of listening at all in the final month. Using an experimental design that separately varies the number of commercial interruptions per hour and the number of ads per commercial interruption, we find that listeners primarily respond to the total number of ads per hour, with a slight preference for more frequent but shorter ad breaks. Lastly, we find that increased ad load led to an increase in the number of paid ad-free subscriptions to Pandora. Importantly, we show that observational methods often lead to biased or even directionally incorrect estimates of these effects, highlighting the value of experimental data.

econ.GN↗

arXiv2021

Disorder Dynamics in Battery Nanoparticles During Phase Transitions Revealed by Operando Single-Particle Diffraction

Structural and ion-ordering phase transitions limit the viability of sodium-ion intercalation materials in grid scale battery storage by reducing their lifetime. However, the combination of phenomena in nanoparticulate electrodes creates complex behavior that is difficult to investigate, especially on the single nanoparticle scale under operating conditions. In this work, operando single-particle x-ray diffraction (oSP-XRD) is used to observe single-particle rotation, interlayer spacing, and layer misorientation in a functional sodium-ion battery. oSP-XRD is applied to Na$_{2/3}$[Ni$_{1/3}$Mn$_{2/3}$]O$_{2}$, an archetypal P2-type sodium-ion positive electrode material with the notorious P2-O2 phase transition induced by sodium (de)intercalation. It is found that during sodium extraction, the misorientation of crystalline layers inside individual particles increases before the layers suddenly align just prior to the P2-O2 transition. The increase in the long-range order coincides with an additional voltage plateau signifying a phase transition prior to the P2-O2 transition. To explain the layer alignment, a model for the phase evolution is proposed that includes a transition from localized to correlated Jahn-Teller distortions. The model is anticipated to guide further characterization and engineering of sodium-ion intercalation materials with P2-O2 type transitions. oSP-XRD therefore opens a powerful avenue for revealing complex phase behavior in heterogeneous nanoparticulate systems.

cond-mat.mtrl-sci↗

arXiv2021

Structure-selective operando x-ray spectroscopy

The relationship between charge and structure dictates the properties of electrochemical systems. For example, reversible Na-ion intercalation - a low-cost alternative to Li-ion technology - often induces detrimental structural phase transformations coupled with charge compensation reactions. However, little is known about the underpinning charge-structure mechanisms because the reduction-oxidation (redox) reactions within coexisting structural phases have so far eluded direct operando investigation. Here, we distinguish x-ray spectra of individual crystalline phases operando during a redox-induced phase transformation in P2-Na2/3Ni1/3Mn2/3O2 - an archetypal layered oxide for sodium-ion batteries. We measure the resonant elastic scattering on the Bragg reflection corresponding to the P2-phase lattice spacing. These resonant spectra become static midway through the sodium extraction in an operando coin cell, while the overall sodium extraction proceeds as evidenced by the X-ray absorption averaging over all electrochemically active Ni atoms. The stop of redox activity in the P2-structure signifies its inability to host Ni4+ ions. The coincident emergence of the O2- structure reveals the rigid link between the local redox and the long-range order during the phase transformation. The structure-selective x-ray spectroscopy thus opens a powerful avenue for resolving the dynamic chemistry of different structural phases in multi-phase electrochemical systems.

cond-mat.mtrl-sci↗

arXiv2020

Designing Differentially Private Estimators in High Dimensions

We study differentially private mean estimation in a high-dimensional setting. Existing differential privacy techniques applied to large dimensions lead to computationally intractable problems or estimators with excessive privacy loss. Recent work in high-dimensional robust statistics has identified computationally tractable mean estimation algorithms with asymptotic dimension-independent error guarantees. We incorporate these results to develop a strict bound on the global sensitivity of the robust mean estimator. This yields a computationally tractable algorithm for differentially private mean estimation in high dimensions with dimension-independent privacy loss. Finally, we show on synthetic data that our algorithm significantly outperforms classic differential privacy methods, overcoming barriers to high-dimensional differential privacy.

cs.LG↗

arXiv2020

Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias

Common methods for interpreting neural models in natural language processing typically examine either their structure or their behavior, but not both. We propose a methodology grounded in the theory of causal mediation analysis for interpreting which parts of a model are causally implicated in its behavior. It enables us to analyze the mechanisms by which information flows from input to output through various model components, known as mediators. We apply this methodology to analyze gender bias in pre-trained Transformer language models. We study the role of individual neurons and attention heads in mediating gender bias across three datasets designed to gauge a model's sensitivity to gender bias. Our mediation analysis reveals that gender bias effects are (i) sparse, concentrated in a small part of the network; (ii) synergistic, amplified or repressed by different components; and (iii) decomposable into effects flowing directly from the input and indirectly through the mediators.

cs.CL↗

arXiv2015

Nanometer scale elemental analysis in the helium ion microscope using time of flight spectrometry

Time of flight backscattering spectrometry (ToF-BS) was successfully implemented in a helium ion microscope (HIM). Its integration introduces the ability to perform laterally resolved elemental analysis as well as elemental depth profiling on the nm scale. A lateral resolution of $\leq$ 54 nm and a time resolution of $\Delta t \leq$ 17 ns $(\Delta t/t \leq 5.4\%)$ are achieved. By using the energy of the backscattered particles for contrast generation, we introduce a new imaging method to the HIM allowing direct elemental mapping as well as local spectrometry. In addition laterally resolved time of flight secondary ion mass spectrometry (ToF-SIMS) can be performed with the same setup. Time of flight is implemented by pulsing the primary ion beam. This is achieved in a cost effective and minimal invasive way that does not influence the high resolution capabilities of the microscope when operating in standard secondary electron (SE) imaging mode. This technique can thus be easily adapted to existing devices. The particular implementation of ToF-BS and ToF-SIMS techniques are described, results are presented and advantages, difficulties and limitations of this new techniques are discussed.

physics.ins-det↗