搜索结果：confirms

The rapidly evolving Node$.$js ecosystem currently includes millions of packages and is a critical part of modern software supply chains, making vulnerability detection of Node$.$js packages increasingly important. However, traditional program analysis struggles in this setting because of dynamic JavaScript features and the large number of package dependencies. Recent advances in large language models (LLMs) and the emerging paradigm of LLM-based agents offer an alternative to handcrafted program models. This raises the question of whether an LLM-centric, tool-augmented approach can effectively detect and confirm taint-style vulnerabilities (e.g., arbitrary command injection) in Node$.$js packages. We implement LLMVD$.$js, a multi-stage agent pipeline to scan code, propose vulnerabilities, generate proof-of-concept exploits, and validate them through lightweight execution oracles; and systematically evaluate its effectiveness in taint-style vulnerability detection and confirmation in Node$.$js packages without dedicated static/dynamic analysis engines for path derivation. For packages from public benchmarks, LLMVD$.$js confirms 84% of the vulnerabilities, compared to less than 22%

This paper studies mechanism design environments in which the designer does not know the distribution of agents' private information a priori and instead learns from agents' behavior induced by the mechanism itself. We formalize a notion of self-confirming mechanisms and a refinement thereof, capturing the idea that an equilibrium mechanism is optimal given the designer's belief and that this belief is consistent with the information produced by the mechanism. We establish a fictitious revelation principle, showing that any incentive-compatible mechanism can be represented as a direct mechanism with filtered type reports that preserve the original mechanism's informational content. Applying the framework to a monopoly problem, we show that, subject to an equilibrium refinement, dominant-strategy self-confirming mechanisms are exactly posted-price mechanisms with locally revenue-maximizing prices.

Combining human and artificial intelligence (AI) is a potentially powerful approach to boost decision accuracy. However, few such approaches exist that effectively integrate both types of intelligence while maintaining human agency. Here, we introduce and evaluate the hybrid confirmation tree, a simple aggregation strategy that compares the independent decisions of both a human and AI, with disagreements triggering a second human tiebreaker. Through analytical derivations, we show that the hybrid confirmation tree can match and exceed the accuracy of a three-person human majority vote while requiring fewer human inputs, particularly when AI accuracy is comparable to or exceeds human accuracy. We analytically demonstrate that the hybrid confirmation tree's ability to achieve complementarity -- outperforming individual humans, AI, and the majority vote -- is maximized when human and AI accuracies are similar and their decisions are not overly correlated. Empirical reanalysis of six real-world datasets (covering skin cancer diagnosis, deepfake detection, geopolitical forecasting, and criminal rearrest) validates these findings, showing that the hybrid confirmation tree improves accura

Context. True spin-orbit obliquities Ψ offer valuable insights into the evolutionary history of exoplanetary systems. Previous studies have suggested that exoplanets tend to occupy either aligned or perpendicular orbits. However, recent research has indicated potential biases caused by the low sample, questioning whether this dichotomy would persist with a larger dataset. Simultaneously, a similar dichotomous behavior has been suggested for Neptune-sized planets. Aims. We aim to investigate the distribution of true spin-orbit obliquities Ψ with an enlarged sample, looking for confirmation of the disputed dichotomy previously found, with a focus also on the obliquities of Neptunes. Methods. Starting from a sample of 264 projected obliquities λ, we homogeneously compute true obliquities Ψ for 116 planets using the rotation period method. We combine them with 4 further values gathered from literature and we then study their distribution, also as a function of various star-planet system parameters. Results. Our data-driven work based on 120 true obliquities Ψ - the largest sample to date - strongly confirms the presence of a single cluster of aligned planets, followed by an isotropic d

Chain-of-thought (CoT) prompting has been widely adopted to enhance the reasoning capabilities of large language models (LLMs). However, the effectiveness of CoT reasoning is inconsistent across tasks with different reasoning types. This work presents a novel perspective to understand CoT behavior through the lens of \textit{confirmation bias} in cognitive psychology. Specifically, we examine how model internal beliefs, approximated by direct question-answering probabilities, affect both reasoning generation ($Q \to R$) and reasoning-guided answer prediction ($QR \to A$) in CoT. By decomposing CoT into a two-stage process, we conduct a thorough correlation analysis in model beliefs, rationale attributes, and stage-wise performance. Our results provide strong evidence of confirmation bias in LLMs, such that model beliefs not only skew the reasoning process but also influence how rationales are utilized for answer prediction. Furthermore, the interplay between task vulnerability to confirmation bias and the strength of beliefs also provides explanations for CoT effectiveness across reasoning tasks and models. Overall, this study provides a valuable insight for the needs of better pro

Confirmation bias, the tendency to interpret information in a way that aligns with one's preconceptions, can profoundly impact scientific research, leading to conclusions that reflect the researcher's hypotheses even when the observational data do not support them. This issue is especially critical in scientific fields involving highly noisy observations, such as cryo-electron microscopy. This study investigates confirmation bias in Gaussian mixture models. We consider the following experiment: A team of scientists assumes they are analyzing data drawn from a Gaussian mixture model with known signals (hypotheses) as centroids. However, in reality, the observations consist entirely of noise without any informative structure. The researchers use a single iteration of the K-means or expectation-maximization algorithms, two popular algorithms to estimate the centroids. Despite the observations being pure noise, we show that these algorithms yield biased estimates that resemble the initial hypotheses, contradicting the unbiased expectation that averaging these noise observations would converge to zero. Namely, the algorithms generate estimates that mirror the postulated model, although

Null Hypothesis Statistical Testing is a dominant framework for conducting statistical analysis across the sciences. There remains considerable debate as to whether, and under what circumstances, evidence can be said to be confirmatory of a null hypothesis. This paper presents a modal logic of short-run frequentist confirmation developed by leveraging the duality between hypothesis testing and statistical estimation. It is shown that a hypothesis is confirmable if and only if it satisfies the topological condition of having nonempty interior. Consequently, two-sided hypotheses are not statistically confirmable owing to defects in their topological structure. Equivalence hypotheses are, by contrast, confirmable.

In introductory physics laboratory instruction, students often expect to confirm or demonstrate textbook physics concepts (Wilcox & Lewandowski, 2017; Hu & Zwickl, 2017; Hu & Zwickl, 2018). This expectation is largely undesirable: labs that emphasize confirmation of textbook physics concepts are unsuccessful at teaching those concepts (Wieman & Holmes, 2015; Holmes et al., 2017) and even in contexts that don't emphasize confirmation, such expectations can lead to students disregarding or manipulating their data in order to obtain the expected result (Smith et al., 2020). In other words, when students expect their lab activities to confirm a known result, they may relinquish epistemic agency and violate disciplinary practices. We claim that, in other cases, confirmatory expectations can actually support productive disciplinary engagement. In particular, when an expected result is not confirmed, students may enter a productive "troubleshooting" mode (Smith et al., 2020). We analyze the complex dynamics of students' epistemological framing in a lab where student's confirmatory expectations support and even generate epistemic agency and disciplinary practices, including

Although resonant planets have orbital periods near commensurability, resonance is also dictated by other factors, such as the planets' eccentricities and masses, and therefore must be confirmed through a study of the system's dynamics. Here, we perform such a study for five multi-planet systems: Kepler-226, Kepler-254, Kepler-363, Kepler-1542, and K2-32. For each system, we run a suite of N-body simulations that span the full parameter-space that is consistent with the constrained orbital and planetary properties. We study the stability of each system and look for resonances based on the libration of the critical resonant angles. We find strong evidence for a two-body resonance in each system; we confirm a 3:2 resonance between Kepler-226c and Kepler-226d, confirm a 3:2 resonance between Kepler-254c and Kepler-254d, and confirm a three-body 1:2:3 resonant chain between the three planets of Kepler-363. We explore the dynamical history of two of these systems and find that these resonances most likely formed without migration. Migration leads to the libration of the three-body resonant angle, but these angles circulate in both Kepler-254 and Kepler-363. Applying our methods to addit

This paper introduces the Conversational Factor Information Retrieval Method (ConFIRM), a novel approach to fine-tuning large language models (LLMs) for domain-specific retrieval tasks. ConFIRM leverages the Five-Factor Model of personality to generate synthetic datasets that accurately reflect target population characteristics, addressing data scarcity in specialized domains. We demonstrate ConFIRM's effectiveness through a case study in the finance sector, fine-tuning a Llama-2-7b model using personality-aligned data from the PolyU-Asklora Fintech Adoption Index. The resulting model achieved 91% accuracy in classifying financial queries, with an average inference time of 0.61 seconds on an NVIDIA A100 GPU. ConFIRM shows promise for creating more accurate and personalized AI-driven information retrieval systems across various domains, potentially mitigating issues of hallucinations and outdated information in LLMs deployed

We report on the astrometric registration of VLBI images of the SiO and water masers in OH231.8+4.2, the iconic Proto-Planetary Nebula also known as the Calabash nebula, using the KVN and Source/Frequency Phase Referencing. This, for the first time, robustly confirms the alignment of the SiO masers, close to the AGB star, driving the bi-lobe structure with the water masers in the out-flow. We are able to trace the bulk motions for the water masers over the last few decades to be 19 km/s and deduce that the age of this expansion stage is 38$\pm$2 years. The combination of this result with the distance allows a full 3D reconstruction, and confirms that the water masers lie on and expand along the known large-scale symmetry axis and that the outflow is only a few decades old, so mass loss is almost certainly on-going. Therefore we conclude that the SiO emission marks the stellar core of the nebular, the water emission traces the expansion, and that there must be multiple epochs of ejection to drive the macro-scale structure.