arXiv subjects

Nima Noorshams

Publications and source records attributed to Nima Noorshams.

10 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↗

arXiv2025

Personalized Interpolation: Achieving Efficient Conversion Estimation with Flexible Optimization Windows

Optimizing conversions is crucial in modern online advertising systems, enabling advertisers to deliver relevant products to users and drive business outcomes. However, accurately predicting conversion events remains challenging due to variable time delays between user interactions (e.g., impressions or clicks) and the actual conversions. These delays vary substantially across advertisers and products, necessitating flexible optimization windows tailored to specific conversion behaviors. To address this, we propose a novel \textit{Personalized Interpolation} method that extends existing models based on fixed conversion windows to support flexible advertiser-specific optimization windows. Our method enables accurate conversion estimation across diverse delay distributions without increasing system complexity. We evaluate the effectiveness of the proposed approach through extensive experiments using a real-world ads conversion model. Our results show that this method achieves both high prediction accuracy and improved efficiency compared to existing solutions. This study demonstrates the potential of our Personalized Interpolation method to improve conversion optimization and support a wider range of advertising strategies in large-scale online advertising systems.

cs.LG↗

arXiv2025

Preference Discerning with LLM-Enhanced Generative Retrieval

In sequential recommendation, models recommend items based on user's interaction history. To this end, current models usually incorporate information such as item descriptions and user intent or preferences. User preferences are usually not explicitly given in open-source datasets, and thus need to be approximated, for example via large language models (LLMs). Current approaches leverage approximated user preferences only during training and rely solely on the past interaction history for recommendations, limiting their ability to dynamically adapt to changing preferences, potentially reinforcing echo chambers. To address this issue, we propose a new paradigm, namely preference discerning, which explicitly conditions a generative recommendation model on user preferences in natural language within its context. To evaluate preference discerning, we introduce a novel benchmark that provides a holistic evaluation across various scenarios, including preference steering and sentiment following. Upon evaluating current state-of-the-art methods on our benchmark, we discover that their ability to dynamically adapt to evolving user preferences is limited. To address this, we propose a new method named Mender ($\textbf{M}$ultimodal Prefer$\textbf{en}$ce $\textbf{D}$iscern$\textbf{er}$), which achieves state-of-the-art performance in our benchmark. Our results show that Mender effectively adapts its recommendation guided by human preferences, even if not observed during training, paving the way toward more flexible recommendation models.

cs.IR↗

arXiv2025

A Unified Knowledge-Distillation and Semi-Supervised Learning Framework to Improve Industrial Ads Delivery Systems

Industrial ads ranking systems conventionally rely on labeled impression data, which leads to challenges such as overfitting, slower incremental gain from model scaling, and biases due to discrepancies between training and serving data. To overcome these issues, we propose a Unified framework for Knowledge-Distillation and Semi-supervised Learning (UKDSL) for ads ranking, empowering the training of models on a significantly larger and more diverse datasets, thereby reducing overfitting and mitigating training-serving data discrepancies. We provide detailed formal analysis and numerical simulations on the inherent miscalibration and prediction bias of multi-stage ranking systems, and show empirical evidence of the proposed framework's capability to mitigate those. Compared to prior work, UKDSL can enable models to learn from a much larger set of unlabeled data, hence, improving the performance while being computationally efficient. Finally, we report the successful deployment of UKDSL in an industrial setting across various ranking models, serving users at multi-billion scale, across various surfaces, geological locations, clients, and optimize for various events, which to the best of our knowledge is the first of its kind in terms of the scale and efficiency at which it operates.

cs.LG↗

arXiv2024

Unifying Generative and Dense Retrieval for Sequential Recommendation

Sequential dense retrieval models utilize advanced sequence learning techniques to compute item and user representations, which are then used to rank relevant items for a user through inner product computation between the user and all item representations. However, this approach requires storing a unique representation for each item, resulting in significant memory requirements as the number of items grow. In contrast, the recently proposed generative retrieval paradigm offers a promising alternative by directly predicting item indices using a generative model trained on semantic IDs that encapsulate items' semantic information. Despite its potential for large-scale applications, a comprehensive comparison between generative retrieval and sequential dense retrieval under fair conditions is still lacking, leaving open questions regarding performance, and computation trade-offs. To address this, we compare these two approaches under controlled conditions on academic benchmarks and propose LIGER (LeveragIng dense retrieval for GEnerative Retrieval), a hybrid model that combines the strengths of these two widely used methods. LIGER integrates sequential dense retrieval into generative retrieval, mitigating performance differences and enhancing cold-start item recommendation in the datasets evaluated. This hybrid approach provides insights into the trade-offs between these approaches and demonstrates improvements in efficiency and effectiveness for recommendation systems in small-scale benchmarks.

cs.IR↗

arXiv2020

TIES: Temporal Interaction Embeddings For Enhancing Social Media Integrity At Facebook

Since its inception, Facebook has become an integral part of the online social community. People rely on Facebook to make connections with others and build communities. As a result, it is paramount to protect the integrity of such a rapidly growing network in a fast and scalable manner. In this paper, we present our efforts to protect various social media entities at Facebook from people who try to abuse our platform. We present a novel Temporal Interaction EmbeddingS (TIES) model that is designed to capture rogue social interactions and flag them for further suitable actions. TIES is a supervised, deep learning, production ready model at Facebook-scale networks. Prior works on integrity problems are mostly focused on capturing either only static or certain dynamic features of social entities. In contrast, TIES can capture both these variant behaviors in a unified model owing to the recent strides made in the domains of graph embedding and deep sequential pattern learning. To show the real-world impact of TIES, we present a few applications especially for preventing spread of misinformation, fake account detection, and reducing ads payment risks in order to enhance the platform's integrity.

cs.AI↗

arXiv2014

A Novel Stochastic Decoding of LDPC Codes with Quantitative Guarantees

Low-density parity-check codes, a class of capacity-approaching linear codes, are particularly recognized for their efficient decoding scheme. The decoding scheme, known as the sum-product, is an iterative algorithm consisting of passing messages between variable and check nodes of the factor graph. The sum-product algorithm is fully parallelizable, owing to the fact that all messages can be update concurrently. However, since it requires extensive number of highly interconnected wires, the fully-parallel implementation of the sum-product on chips is exceedingly challenging. Stochastic decoding algorithms, which exchange binary messages, are of great interest for mitigating this challenge and have been the focus of extensive research over the past decade. They significantly reduce the required wiring and computational complexity of the message-passing algorithm. Even though stochastic decoders have been shown extremely effective in practice, the theoretical aspect and understanding of such algorithms remains limited at large. Our main objective in this paper is to address this issue. We first propose a novel algorithm referred to as the Markov based stochastic decoding. Then, we provide concrete quantitative guarantees on its performance for tree-structured as well as general factor graphs. More specifically, we provide upper-bounds on the first and second moments of the error, illustrating that the proposed algorithm is an asymptotically consistent estimate of the sum-product algorithm. We also validate our theoretical predictions with experimental results, showing we achieve comparable performance to other practical stochastic decoders.

cs.IT↗

arXiv2013

Non-Asymptotic Analysis of an Optimal Algorithm for Network-Constrained Averaging with Noisy Links

The problem of network-constrained averaging is to compute the average of a set of values distributed throughout a graph G using an algorithm that can pass messages only along graph edges. We study this problem in the noisy setting, in which the communication along each link is modeled by an additive white Gaussian noise channel. We propose a two-phase decentralized algorithm, and we use stochastic approximation methods in conjunction with the spectral graph theory to provide concrete (non-asymptotic) bounds on the mean-squared error. Having found such bounds, we analyze how the number of iterations T_G(n; δ) required to achieve mean-squared error δ scales as a function of the graph topology and the number of nodes n. Previous work provided guarantees with the number of iterations scaling inversely with the second smallest eigenvalue of the Laplacian. This paper gives an algorithm that reduces this graph dependence to the graph diameter, which is the best scaling possible.

cs.DC↗

arXiv2012

Belief Propagation for Continuous State Spaces: Stochastic Message-Passing with Quantitative Guarantees

The sum-product or belief propagation (BP) algorithm is a widely used message-passing technique for computing approximate marginals in graphical models. We introduce a new technique, called stochastic orthogonal series message-passing (SOSMP), for computing the BP fixed point in models with continuous random variables. It is based on a deterministic approximation of the messages via orthogonal series expansion, and a stochastic approximation via Monte Carlo estimates of the integral updates of the basis coefficients. We prove that the SOSMP iterates converge to a δ-neighborhood of the unique BP fixed point for any tree-structured graph, and for any graphs with cycles in which the BP updates satisfy a contractivity condition. In addition, we demonstrate how to choose the number of basis coefficients as a function of the desired approximation accuracy δand smoothness of the compatibility functions. We illustrate our theory with both simulated examples and in application to optical flow estimation.

cs.IT↗

arXiv2012

Stochastic Belief Propagation: A Low-Complexity Alternative to the Sum-Product Algorithm

The sum-product or belief propagation (BP) algorithm is a widely-used message-passing algorithm for computing marginal distributions in graphical models with discrete variables. At the core of the BP message updates, when applied to a graphical model with pairwise interactions, lies a matrix-vector product with complexity that is quadratic in the state dimension $d$, and requires transmission of a $(d-1)$-dimensional vector of real numbers (messages) to its neighbors. Since various applications involve very large state dimensions, such computation and communication complexities can be prohibitively complex. In this paper, we propose a low-complexity variant of BP, referred to as stochastic belief propagation (SBP). As suggested by the name, it is an adaptively randomized version of the BP message updates in which each node passes randomly chosen information to each of its neighbors. The SBP message updates reduce the computational complexity (per iteration) from quadratic to linear in $d$, without assuming any particular structure of the potentials, and also reduce the communication complexity significantly, requiring only $\log{d}$ bits transmission per edge. Moreover, we establish a number of theoretical guarantees for the performance of SBP, showing that it converges almost surely to the BP fixed point for any tree-structured graph, and for graphs with cycles satisfying a contractivity condition. In addition, for these graphical models, we provide non-asymptotic upper bounds on the convergence rate, showing that the $\ell_{\infty}$ norm of the error vector decays no slower than $O(1/\sqrt{t})$ with the number of iterations $t$ on trees and the mean square error decays as $O(1/t)$ for general graphs. These analysis show that SBP can provably yield reductions in computational and communication complexities for various classes of graphical models.

cs.IT↗