arXiv subjects

Christopher Jung

Publications and source records attributed to Christopher Jung.

At least 19 recordsLinked to original sources

arXiv2026

Convex Dataset Valuation for Post-Training

Improving LLM performance on downstream tasks sometimes requires leveraging auxiliary datasets during post-training. In practice, however, developers face constraints on compute, labeling, and licensing costs that preclude using all available data, necessitating principled dataset-level selection. These constraints are increasingly shaped by dataset marketplaces, where data acquisition is governed by budgets and negotiation. We study dataset valuation as a subset selection problem during LLM post-training. Our goal is to identify and weight auxiliary datasets so as to maximize target task performance given constrained budgets. We first show that commonly used gradient alignment scores provide a reasonable yet incomplete valuation signal, as they ignore redundancy among datasets. To address this, we propose a scalable convex dataset-level valuation method based on kernel mean matching (KMM) in gradient space, which jointly accounts for alignment with the target task and redundancy across auxiliary datasets. Through extensive experiments across diverse post-training settings and tasks, we show that our approach consistently outperforms existing valuation baselines, achieving stronger performance with low computational overhead. Our results position dataset valuation as a practical decision tool for post-training data selection in market-constrained large language model settings. The code is available at https://github.com/uiuctml/convex_data_valuation.

cs.LG↗

arXiv2026

Expectation Error Bounds for Transfer Learning in Linear Regression and Linear Neural Networks

In transfer learning, the learner leverages auxiliary data to improve generalization on a main task. However, the precise theoretical understanding of when and how auxiliary data help remains incomplete. We provide new insights on this issue in two canonical linear settings: ordinary least squares regression and under-parameterized linear neural networks. For linear regression, we derive exact closed-form expressions for the expected generalization error with bias-variance decomposition, yielding necessary and sufficient conditions for auxiliary tasks to improve generalization on the main task. We also derive globally optimal task weights as outputs of solvable optimization programs, with consistency guarantees for empirical estimates. For linear neural networks with shared representations of width $q \leq K$, where $K$ is the number of auxiliary tasks, we derive a non-asymptotic expectation bound on the generalization error, yielding the first non-vacuous sufficient condition for beneficial auxiliary learning in this setting, as well as principled directions for task weight curation. We achieve this by proving a new column-wise low-rank perturbation bound for random matrices, which improves upon existing bounds by preserving fine-grained column structures. Our results are verified on synthetic data simulated with controlled parameters.

cs.LG↗

arXiv2025

Cross-attention Secretly Performs Orthogonal Alignment in Recommendation Models

Cross-domain sequential recommendation (CDSR) aims to align heterogeneous user behavior sequences collected from different domains. While cross-attention is widely used to enhance alignment and improve recommendation performance, its underlying mechanism is not fully understood. Most researchers interpret cross-attention as residual alignment, where the output is generated by removing redundant and preserving non-redundant information from the query input by referencing another domain data which is input key and value. Beyond the prevailing view, we introduce Orthogonal Alignment, a phenomenon in which cross-attention discovers novel information that is not present in the query input, and further argue that those two contrasting alignment mechanisms can co-exist in recommendation models We find that when the query input and output of cross-attention are orthogonal, model performance improves over 300 experiments. Notably, Orthogonal Alignment emerges naturally, without any explicit orthogonality constraints. Our key insight is that Orthogonal Alignment emerges naturally because it improves scaling law. We show that baselines additionally incorporating cross-attention module outperform parameter-matched baselines, achieving a superior accuracy-per-model parameter. We hope these findings offer new directions for parameter-efficient scaling in multi-modal research.

cs.LG↗

arXiv2025

Novel Numerical Method for Simultaneous Design and Control Optimization of Seasonal Thermal Energy Storage Systems

The transition to a carbon-neutral energy system requires widespread deployment of renewable energy sources and economically feasible energy storage solutions. This study presents a comprehensive optimization framework that jointly addresses the design and control of a nonlinear energy system supplying both heat and electricity to the Dietenbach district in Freiburg, Germany. The proposed system integrates solar and wind power with battery storage and seasonal thermal energy storage coupled via a heat pump, enhancing self-sufficiency and mitigating seasonal supply-demand mismatches. A multi-node lumped-parameter model captures heat transfer within the pit thermal energy storage, forming the basis of a periodic optimal control problem solved numerically. An averaging method reduces computation time by 80.5% while preserving fidelity for year-long optimization. A case study shows a projected total yearly energy cost of 5.93 EUR/m2 for combined heat and electricity, which is 73% lower than the German average. This study underscores the feasibility of designing economically viable, autonomous energy communities in real-world scenarios and provides an efficient, robust optimization framework for designing system components and operational control strategies.

math.OC↗

arXiv2024

Orthogonal Causal Calibration

Estimates of heterogeneous treatment effects such as conditional average treatment effects (CATEs) and conditional quantile treatment effects (CQTEs) play an important role in real-world decision making. Given this importance, one should ensure these estimates are calibrated. While there is a rich literature on calibrating estimators of non-causal parameters, very few methods have been derived for calibrating estimators of causal parameters, or more generally estimators of quantities involving nuisance parameters. In this work, we develop general algorithms for reducing the task of causal calibration to that of calibrating a standard (non-causal) predictive model. Throughout, we study a notion of calibration defined with respect to an arbitrary, nuisance-dependent loss $\ell$, under which we say an estimator $\theta$ is calibrated if its predictions cannot be changed on any level set to decrease loss. For losses $\ell$ satisfying a condition called universal orthogonality, we present a simple algorithm that transforms partially-observed data into generalized pseudo-outcomes and applies any off-the-shelf calibration procedure. For losses $\ell$ satisfying a weaker assumption called conditional orthogonality, we provide a similar sample splitting algorithm the performs empirical risk minimization over an appropriately defined class of functions. Convergence of both algorithms follows from a generic, two term upper bound of the calibration error of any model. We demonstrate the practical applicability of our results in experiments on both observational and synthetic data. Our results are exceedingly general, showing that essentially any existing calibration algorithm can be used in causal settings, with additional loss only arising from errors in nuisance estimation.

stat.ML↗

arXiv2024

Taking a Moment for Distributional Robustness

A rich line of recent work has studied distributionally robust learning approaches that seek to learn a hypothesis that performs well, in the worst-case, on many different distributions over a population. We argue that although the most common approaches seek to minimize the worst-case loss over distributions, a more reasonable goal is to minimize the worst-case distance to the true conditional expectation of labels given each covariate. Focusing on the minmax loss objective can dramatically fail to output a solution minimizing the distance to the true conditional expectation when certain distributions contain high levels of label noise. We introduce a new min-max objective based on what is known as the adversarial moment violation and show that minimizing this objective is equivalent to minimizing the worst-case $\ell_2$-distance to the true conditional expectation if we take the adversary's strategy space to be sufficiently rich. Previous work has suggested minimizing the maximum regret over the worst-case distribution as a way to circumvent issues arising from differential noise levels. We show that in the case of square loss, minimizing the worst-case regret is also equivalent to minimizing the worst-case $\ell_2$-distance to the true conditional expectation. Although their objective and our objective both minimize the worst-case distance to the true conditional expectation, we show that our approach provides large empirical savings in computational cost in terms of the number of groups, while providing the same noise-oblivious worst-distribution guarantee as the minimax regret approach, thus making positive progress on an open question posed by Agarwal and Zhang (2022).

cs.LG↗

arXiv2023

Oracle Efficient Online Multicalibration and Omniprediction

A recent line of work has shown a surprising connection between multicalibration, a multi-group fairness notion, and omniprediction, a learning paradigm that provides simultaneous loss minimization guarantees for a large family of loss functions. Prior work studies omniprediction in the batch setting. We initiate the study of omniprediction in the online adversarial setting. Although there exist algorithms for obtaining notions of multicalibration in the online adversarial setting, unlike batch algorithms, they work only for small finite classes of benchmark functions $F$, because they require enumerating every function $f \in F$ at every round. In contrast, omniprediction is most interesting for learning theoretic hypothesis classes $F$, which are generally continuously large. We develop a new online multicalibration algorithm that is well defined for infinite benchmark classes $F$, and is oracle efficient (i.e. for any class $F$, the algorithm has the form of an efficient reduction to a no-regret learning algorithm for $F$). The result is the first efficient online omnipredictor -- an oracle efficient prediction algorithm that can be used to simultaneously obtain no regret guarantees to all Lipschitz convex loss functions. For the class $F$ of linear functions, we show how to make our algorithm efficient in the worst case. Also, we show upper and lower bounds on the extent to which our rates can be improved: our oracle efficient algorithm actually promises a stronger guarantee called swap-omniprediction, and we prove a lower bound showing that obtaining $O(\sqrt{T})$ bounds for swap-omniprediction is impossible in the online setting. On the other hand, we give a (non-oracle efficient) algorithm which can obtain the optimal $O(\sqrt{T})$ omniprediction bounds without going through multicalibration, giving an information theoretic separation between these two solution concepts.

cs.LG↗

arXiv2022

Batch Multivalid Conformal Prediction

We develop fast distribution-free conformal prediction algorithms for obtaining multivalid coverage on exchangeable data in the batch setting. Multivalid coverage guarantees are stronger than marginal coverage guarantees in two ways: (1) They hold even conditional on group membership -- that is, the target coverage level $1-\alpha$ holds conditionally on membership in each of an arbitrary (potentially intersecting) group in a finite collection $\mathcal{G}$ of regions in the feature space. (2) They hold even conditional on the value of the threshold used to produce the prediction set on a given example. In fact multivalid coverage guarantees hold even when conditioning on group membership and threshold value simultaneously. We give two algorithms: both take as input an arbitrary non-conformity score and an arbitrary collection of possibly intersecting groups $\mathcal{G}$, and then can equip arbitrary black-box predictors with prediction sets. Our first algorithm (BatchGCP) is a direct extension of quantile regression, needs to solve only a single convex minimization problem, and produces an estimator which has group-conditional guarantees for each group in $\mathcal{G}$. Our second algorithm (BatchMVP) is iterative, and gives the full guarantees of multivalid conformal prediction: prediction sets that are valid conditionally both on group membership and non-conformity threshold. We evaluate the performance of both of our algorithms in an extensive set of experiments. Code to replicate all of our experiments can be found at https://github.com/ProgBelarus/BatchMultivalidConformal

cs.LG↗

arXiv2022

Multicalibrated Regression for Downstream Fairness

We show how to take a regression function $\hat{f}$ that is appropriately ``multicalibrated'' and efficiently post-process it into an approximately error minimizing classifier satisfying a large variety of fairness constraints. The post-processing requires no labeled data, and only a modest amount of unlabeled data and computation. The computational and sample complexity requirements of computing $\hat f$ are comparable to the requirements for solving a single fair learning task optimally, but it can in fact be used to solve many different downstream fairness-constrained learning problems efficiently. Our post-processing method easily handles intersecting groups, generalizing prior work on post-processing regression functions to satisfy fairness constraints that only applied to disjoint groups. Our work extends recent work showing that multicalibrated regression functions are ``omnipredictors'' (i.e. can be post-processed to optimally solve unconstrained ERM problems) to constrained optimization.

cs.LG↗

arXiv2022

Practical Adversarial Multivalid Conformal Prediction

We give a simple, generic conformal prediction method for sequential prediction that achieves target empirical coverage guarantees against adversarially chosen data. It is computationally lightweight -- comparable to split conformal prediction -- but does not require having a held-out validation set, and so all data can be used for training models from which to derive a conformal score. It gives stronger than marginal coverage guarantees in two ways. First, it gives threshold calibrated prediction sets that have correct empirical coverage even conditional on the threshold used to form the prediction set from the conformal score. Second, the user can specify an arbitrary collection of subsets of the feature space -- possibly intersecting -- and the coverage guarantees also hold conditional on membership in each of these subsets. We call our algorithm MVP, short for MultiValid Prediction. We give both theory and an extensive set of empirical evaluations.

cs.LG↗

arXiv2022

Distributionally Robust Data Join

Suppose we are given two datasets: a labeled dataset and unlabeled dataset which also has additional auxiliary features not present in the first dataset. What is the most principled way to use these datasets together to construct a predictor? The answer should depend upon whether these datasets are generated by the same or different distributions over their mutual feature sets, and how similar the test distribution will be to either of those distributions. In many applications, the two datasets will likely follow different distributions, but both may be close to the test distribution. We introduce the problem of building a predictor which minimizes the maximum loss over all probability distributions over the original features, auxiliary features, and binary labels, whose Wasserstein distance is $r_1$ away from the empirical distribution over the labeled dataset and $r_2$ away from that of the unlabeled dataset. This can be thought of as a generalization of distributionally robust optimization (DRO), which allows for two data sources, one of which is unlabeled and may contain auxiliary features.

cs.LG↗

arXiv2022

Getting more with less? Why repowering onshore wind farms does not always lead to more wind power generation -- a German case study

The best wind locations are nowadays often occupied by old, less efficient and relatively small wind turbines. Many of them will soon reach the end of their operating lifetime, or lose financial support. Therefore, repowering comes to the fore. However, social acceptance and land use restrictions have been under constant change since the initial expansions, which makes less area available for new turbines, even on existing sites. For the example of Germany, this study assesses the repowering potential for onshore wind energy in high detail, on the basis of regionally differentiated land eligibility criteria. The results show that under the given regional criteria, repowering will decrease both operating capacity and annual energy yield by roughly 40\,\% compared to the status quo. This is because around half of the wind turbines are currently located in restricted areas, given newly enacted exclusion criteria. Sensitivity analyses on the exclusion criteria show that the minimum distance to discontinuous urban fabric is the most sensitive criterion in determining the number of turbines that can be repowered. As regulations on this can vary substantially across different regions, the location-specific methodology chosen here can assess the repowering potential more realistically than existing approaches.

physics.soc-ph↗

arXiv2021

Self-consistent $O(4)$ model spectral functions from analytically continued FRG flows

In this paper we explore practicable ways for self-consistent calculations of spectral functions from analytically continued functional renormalization group (aFRG) flow equations. As a particularly straightforward one we propose to include parametrizations of self-energies based on explicit analytic one-loop expressions. To exemplify this scheme we calculate the spectral functions of pion and sigma meson of the $O(4)$ model at vanishing temperature in the broken phase. Comparing the results with those from previous aFRG calculations, we explicitly demonstrate how self-consistency at all momenta fixes the tight relation between particle masses and decay thresholds. In addition, the two-point functions from our new semi-analytic FRG scheme have the desired domain of holomorphy built in and can readily be studied in the entire cut-complex frequency plane, on physical as well as other Riemann sheets. This is very illustrative and allows, for example, to trace the flow of the resonance pole of the sigma meson across an unphysical sheet. In order to assess the limitations due to the underlying one-loop structure, we also introduce a fully self-consistent numerical scheme based on spectral representations with scale-dependent spectral functions. The most notable improvement of this numerically involved calculation is that it describes the three-particle resonance decay of an off-shell pion, $\pi^* \to \sigma\pi\to3\pi$. Apart from this further conceptual improvement, overall agreement with the results from the considerably simpler semi-analytic one-loop scheme is very encouraging, however. The latter can therefore provide a sound and practicable basis for self-consistent calculations of spectral functions in more realistic effective theories for warm and dense matter.

hep-ph↗

arXiv2021

Adaptive Machine Unlearning

Data deletion algorithms aim to remove the influence of deleted data points from trained models at a cheaper computational cost than fully retraining those models. However, for sequences of deletions, most prior work in the non-convex setting gives valid guarantees only for sequences that are chosen independently of the models that are published. If people choose to delete their data as a function of the published models (because they don't like what the models reveal about them, for example), then the update sequence is adaptive. In this paper, we give a general reduction from deletion guarantees against adaptive sequences to deletion guarantees against non-adaptive sequences, using differential privacy and its connection to max information. Combined with ideas from prior work which give guarantees for non-adaptive deletion sequences, this leads to extremely flexible algorithms able to handle arbitrary model classes and training methodologies, giving strong provable deletion guarantees for adaptive deletion sequences. We show in theory how prior work for non-convex models fails against adaptive deletion sequences, and use this intuition to design a practical attack against the SISA algorithm of Bourtoule et al. [2021] on CIFAR-10, MNIST, Fashion-MNIST.

cs.LG↗

arXiv2021

Vector and Axial-Vector Mesons in Nuclear Matter

As a first step towards a realistic phenomenological description of vector and axial-vector mesons in nuclear matter, we calculate the spectral functions of the $\rho$ and the $a_1$ meson in a chiral baryon-meson model as a low-energy effective realization of QCD, taking into account the effects of fluctuations from scalar mesons, nucleons, and vector mesons within the Functional Renormalization Group (FRG) approach. The phase diagram of the effective hadronic theory exhibits a nuclear liquid-gas phase transition as well as a chiral phase transition at a higher baryon-chemical potential. The in-medium $\rho$ and $a_1$ spectral functions are calculated by using the previously introduced analytically-continued FRG (aFRG) method. Our results show strong modifications of the spectral functions in particular near the critical endpoints of both phase transitions which may well be of relevance for electromagnetic rates in heavy-ion collisions or neutrino emissivities in neutron-star merger events.

hep-ph↗

arXiv2021

Online Multivalid Learning: Means, Moments, and Prediction Intervals

We present a general, efficient technique for providing contextual predictions that are "multivalid" in various senses, against an online sequence of adversarially chosen examples $(x,y)$. This means that the resulting estimates correctly predict various statistics of the labels $y$ not just marginally -- as averaged over the sequence of examples -- but also conditionally on $x \in G$ for any $G$ belonging to an arbitrary intersecting collection of groups $\mathcal{G}$. We provide three instantiations of this framework. The first is mean prediction, which corresponds to an online algorithm satisfying the notion of multicalibration from Hebert-Johnson et al. The second is variance and higher moment prediction, which corresponds to an online algorithm satisfying the notion of mean-conditioned moment multicalibration from Jung et al. Finally, we define a new notion of prediction interval multivalidity, and give an algorithm for finding prediction intervals which satisfy it. Because our algorithms handle adversarially chosen examples, they can equally well be used to predict statistics of the residuals of arbitrary point prediction methods, giving rise to very general techniques for quantifying the uncertainty of predictions of black box algorithms, even in an online adversarial setting. When instantiated for prediction intervals, this solves a similar problem as conformal prediction, but in an adversarial environment and with multivalidity guarantees stronger than simple marginal coverage guarantees.

cs.LG↗

arXiv2020

An Algorithmic Framework for Fairness Elicitation

We consider settings in which the right notion of fairness is not captured by simple mathematical definitions (such as equality of error rates across groups), but might be more complex and nuanced and thus require elicitation from individual or collective stakeholders. We introduce a framework in which pairs of individuals can be identified as requiring (approximately) equal treatment under a learned model, or requiring ordered treatment such as "applicant Alice should be at least as likely to receive a loan as applicant Bob". We provide a provably convergent and oracle efficient algorithm for learning the most accurate model subject to the elicited fairness constraints, and prove generalization bounds for both accuracy and fairness. This algorithm can also combine the elicited constraints with traditional statistical fairness notions, thus "correcting" or modifying the latter by the former. We report preliminary findings of a behavioral study of our framework using human-subject fairness constraints elicited on the COMPAS criminal recidivism dataset.

cs.LG↗

arXiv2020

Moment Multicalibration for Uncertainty Estimation

We show how to achieve the notion of "multicalibration" from Hébert-Johnson et al. [2018] not just for means, but also for variances and other higher moments. Informally, it means that we can find regression functions which, given a data point, can make point predictions not just for the expectation of its label, but for higher moments of its label distribution as well-and those predictions match the true distribution quantities when averaged not just over the population as a whole, but also when averaged over an enormous number of finely defined subgroups. It yields a principled way to estimate the uncertainty of predictions on many different subgroups-and to diagnose potential sources of unfairness in the predictive power of features across subgroups. As an application, we show that our moment estimates can be used to derive marginal prediction intervals that are simultaneously valid as averaged over all of the (sufficiently large) subgroups for which moment multicalibration has been obtained.

cs.LG↗