arXiv subjects

Aymen Echarghaoui

Publications and source records attributed to Aymen Echarghaoui.

4 recordsLinked to original sources

arXiv2026

BALAR : A Bayesian Agentic Loop for Active Reasoning

Large language models increasingly operate in interactive settings where solving a task requires multiple rounds of information exchange with a user. However, most current systems treat dialogue reactively and lack a principled mechanism to reason about what information is missing and which question should be asked next. We propose BALAR (Bayesian Agentic Loop for Active Reasoning), a task-agnostic outer-loop algorithm that requires no fine-tuning and enables structured multi-turn interaction between an LLM agent and a user. BALAR maintains a structured belief over latent states, selects clarifying questions by maximizing expected mutual information, and dynamically expands its state representation when the current one proves insufficient. We evaluate BALAR on three diverse benchmarks: AR-Bench-DC (detective cases), AR-Bench-SP (thinking puzzles), and iCraft-MD (clinical diagnosis). BALAR significantly outperforms all baselines across all three benchmarks, with $14.6\%$ higher accuracy on AR-Bench-DC, $38.5\%$ on AR-Bench-SP, and $30.5\%$ on iCraft-MD.

cs.AI↗

arXiv2026

Univariate-Guided Interaction Modeling

We propose a procedure for sparse regression with pairwise interactions, by generalizing the Univariate Guided Sparse Regression (UniLasso) methodology. A central contribution is our introduction of a concept of univariate (or marginal) interactions. Using this concept, we propose two algorithms -- uniPairs and uniPairs-2stage -- , and evaluate their performance against established methods, including Glinternet and Sprinter. We show that our framework yields sparser models with more interpretable interactions. We also prove support recovery results for our proposal under suitable conditions.

stat.ME↗

arXiv2025

Intelligently Weighting Multiple Reference Models for Direct Preference Optimization of LLMs

Fine-tuning is integral for aligning large language models (LLMs) with human preferences. Multiple-Reference Preference Optimization (MRPO) builds on Direct Preference Optimization (DPO) by fine-tuning LLMs on preference datasets while regularizing the policy towards a mixture of reference models to leverage their collective desirable properties. However, current methods for setting the reference weights are ad-hoc and statistically unsound, leading to unreliable performance. To address this, we introduce four new weighting strategies: two offline methods that leverage held-out validation signal; one online method that uses a sliding-window estimator to reduce overfitting; and an online method that treats reference weighting as a $K$-armed bandit via Thompson Sampling. Experiments using Qwen2.5-0.5B as the policy model and seven reference models from the Llama, Mistral, Qwen, Yi, and Phi families (0.5B-14B each) show that all 4 of our strategies outperform the current MRPO weighting methods on UltraFeedback and SafeRLHF in preference accuracy. More thought-provokingly, however, we find that single-reference DPO, using any of 6 out of 7 references, consistently outperforms all tested multiple-reference approaches -- calling into question the practical appeal of multiple-reference approaches.

cs.LG↗

arXiv2023

Lasting Diversity and Superior Runtime Guarantees for the $(μ+1)$ Genetic Algorithm

Most evolutionary algorithms (EAs) used in practice employ crossover. In contrast, only for few and mostly artificial examples a runtime advantage from crossover could be proven with mathematical means. The most convincing such result shows that the $(μ+1)$ genetic algorithm (GA) with population size $μ=O(n)$ optimizes jump functions with gap size $k \ge 3$ in time $O(n^k / μ+ n^{k-1}\log n)$, beating the $Θ(n^k)$ runtime of many mutation-based EAs. This result builds on a proof that the GA occasionally and then for an expected number of $Ω(μ^2)$ iterations has a population that is not dominated by a single genotype. In this work, we show that this diversity persist with high probability for a time exponential in $μ$ (instead of quadratic). From this better understanding of the population diversity, we obtain stronger runtime guarantees, among them the statement that for all $c\ln(n)\leμ\le n/\log n$, with $c$ a suitable constant, the runtime of the $(μ+1)$ GA on $\mathrm{Jump}_k$, with $k \ge 3$, is $O(n^{k-1})$. Consequently, already with logarithmic population sizes, the GA gains a speed-up of order $Ω(n)$ from crossover.

cs.NE↗