arXiv:2601.05219v2 Announce Type: replace Abstract: One-shot prediction enables rapid adaptation of pretrained foundation models to new tasks using only one labeled example, but lacks principled uncertainty quantification. While conformal prediction provides finite-sample coverage guarantees, standard split conformal methods are inefficient in the one-shot setting due to data splitting and reliance on a single predictor. We propose Conformal Aggregation of One-Shot Predictors (CAOS), a conformal framework that adaptively aggregates multiple one-shot predictors and uses a leave-one-out calibration scheme to […]

Ver mais

arXiv:2601.00231v1 Announce Type: new Abstract: Parameter-efficient fine-tuning (PEFT) is the default way to adapt LLMs, but widely used LoRA and QLoRA are largely geometry-agnostic: they optimize in fixed, randomly oriented low-rank subspaces with first-order descent, mostly ignoring local loss curvature. This can inflate the effective update budget and amplify drift along weakly constrained directions. We introduce GRIT, a dynamic, curvature-aware LoRA procedure that preserves the LoRA parameterization but: (1) preconditions gradients in rank space using K-FAC as a […]

Ver mais

arXiv:2604.22757v1 Announce Type: new Abstract: We introduce StratRAG, an open-source retrieval evaluation dataset for benchmarking Retrieval-Augmented Generation (RAG) systems on multi-hop reasoning tasks under realistic, noisy document-pool conditions. Derived from HotpotQA (distractor setting), StratRAG comprises 2,200 examples across three question types — bridge, comparison, and yes-no — each paired with a pool of 15 candidate documents containing exactly 2 gold documents and 13 topically related distractors. We benchmark three retrieval strategies — BM25, dense retrieval (all-MiniLM-L6-v2), and hybrid […]

Ver mais

arXiv:2602.16810v1 Announce Type: new Abstract: Unlike in TFNP, for which there is an abundance of problems capturing natural existence principles which are incomparable (in the black-box setting), Kleinberg et al. [KKMP21] observed that many of the natural problems considered so far in the second level of the total function polynomial hierarchy (TF$Sigma_2$) reduce to the Strong Avoid problem. In this work, we prove that the Linear Ordering Principle does not reduce to Strong Avoid in the black-box setting, […]

Ver mais

There is a growing demand for efficient data removal to comply with regulations like the GDPR and to mitigate the influence of biased or corrupted data. This has motivated the field of machine unlearning, which aims to eliminate the influence of specific data subsets without the cost of full retraining. In this work, we propose a statistical framework for machine unlearning with generic loss functions and establish theoretical guarantees. For squared loss, especially, we develop Unlearning Least Squares […]

Ver mais

arXiv:2509.16180v2 Announce Type: replace-cross Abstract: We propose an algorithm with improved query-complexity for the problem of hypothesis selection under local differential privacy constraints. Given a set of $k$ probability distributions $Q$, we describe an algorithm that satisfies local differential privacy, performs $tilde{O}(k^{3/2})$ non-adaptive queries to individuals who each have samples from a probability distribution $p$, and outputs a probability distribution from the set $Q$ which is nearly the closest to $p$. Previous algorithms required either $Omega(k^2)$ queries or […]

Ver mais

arXiv:2604.07364v1 Announce Type: new Abstract: Alphanumeric identifiers such as manufacturer part numbers (MPNs), SKUs, and model codes are ubiquitous in e-commerce catalogs and search. These identifiers are sparse, non linguistic, and highly sensitive to tokenization and typographical variation, rendering conventional lexical and embedding based retrieval methods ineffective. We propose a training free, character level retrieval framework that encodes each alphanumeric sequence as a fixed length binary vector. This representation enables efficient similarity computation via Hamming distance and supports […]

Ver mais

arXiv:2603.03322v1 Announce Type: new Abstract: Recent advancements in Large Language Model (LLM) agents have demonstrated remarkable potential in automatic knowledge discovery. However, rigorously evaluating an AI’s capacity for knowledge discovery remains a critical challenge. Existing benchmarks predominantly rely on static datasets, leading to inevitable data contamination where models have likely seen the evaluation knowledge during training. Furthermore, the rapid release cycles of modern LLMs render static benchmarks quickly outdated, failing to assess the ability to discover truly new […]

Ver mais

OpenAI on just launched Daybreak, a cybersecurity initiative that combines the company’s frontier AI models with Codex Security, its coding-focused agentic system, and a broad network of security partners. The program is aimed at developers, enterprise security teams, researchers, and government-linked defenders who need to find, validate, and patch software vulnerabilities earlier in the development cycle — not after exploits have already been identified in the wild. The core premise of Daybreak is a shift in how software […]

Ver mais