搜索结果：experimenting

We provide an overview of methods for designing and implementing experiments (field, lab, hybrid, and natural) when there are networks of interactions between subjects.

Online social networks have transformed the ways in which political mobilization messages are disseminated, raising new questions about how peer influence operates at scale. Building on the landmark 61-million-person Facebook experiment \citep{bond201261}, we develop an agent-based simulation framework that integrates real U.S. Census demographic distributions, authentic Twitter network topology, and heterogeneous large language model (LLM) agents to examine the effect of mobilization messages on voter turnout. Each simulated agent is assigned demographic attributes, a personal political stance, and an LLM variant (\texttt{GPT-4.1}, \texttt{GPT-4.1-Mini}, or \texttt{GPT-4.1-Nano}) reflecting its political sophistication. Agents interact over realistic social network structures, receiving personalized feeds and dynamically updating their engagement behaviors and voting intentions. Experimental conditions replicate the informational and social mobilization treatments of the original Facebook study. Across scenarios, the simulator reproduces qualitative patterns observed in field experiments, including stronger mobilization effects under social message treatments and measurable peer s

Online experiments in internet systems, also known as A/B tests, are used for a wide range of system tuning problems, such as optimizing recommender system ranking policies and learning adaptive streaming controllers. Decision-makers generally wish to optimize for long-term treatment effects of the system changes, which often requires running experiments for a long time as short-term measurements can be misleading due to non-stationarity in treatment effects over time. The sequential experimentation strategies--which typically involve several iterations--can be prohibitively long in such cases. We describe a novel approach that combines fast experiments (e.g., biased experiments run only for a few hours or days) and/or offline proxies (e.g., off-policy evaluation) with long-running, slow experiments to perform sequential, Bayesian optimization over large action spaces in a short amount of time.

Large language models (LLMs) are powerful but resource intensive, limiting accessibility. HITgram addresses this gap by offering a lightweight platform for n-gram model experimentation, ideal for resource-constrained environments. It supports unigrams to 4-grams and incorporates features like context sensitive weighting, Laplace smoothing, and dynamic corpus management to e-hance prediction accuracy, even for unseen word sequences. Experiments demonstrate HITgram's efficiency, achieving 50,000 tokens/second and generating 2-grams from a 320MB corpus in 62 seconds. HITgram scales efficiently, constructing 4-grams from a 1GB file in under 298 seconds on an 8 GB RAM system. Planned enhancements include multilingual support, advanced smoothing, parallel processing, and model saving, further broadening its utility.

This paper builds on top of a paper we have published very recently, in which we have proposed a novel approach to prime factorization (PF) by quantum annealing, where 8,219,999=32,749x251 was the highest prime product we were able to factorize -- which, to the best of our knowledge is the largest number which was ever factorized by means of a quantum device. The series of annealing experiments which led us to these results, however, did not follow a straight-line path; rather, they involved a convoluted trial-and-error process, full of failed or partially-failed attempts and backtracks, which only in the end drove us to find the successful annealing strategies. In this paper, we delve into the reasoning behind our experimental decisions and provide an account of some of the attempts we have taken before conceiving the final strategies that allowed us to achieve the results. This involves also a bunch of ideas, techniques, and strategies we investigated which, although turned out to be inferior wrt. those we adopted in the end, may instead provide insights to a more-specialized audience of D-Wave users and practitioners. In particular, we show the following insights: ($i$) differen

Interactive Virtual Reality (VR) streaming over Wi-Fi networks encounters significant challenges due to bandwidth fluctuations caused by channel contention and user mobility. Adaptive BitRate (ABR) algorithms dynamically adjust the video encoding bitrate based on the available network capacity, aiming to maximize image quality while mitigating congestion and preserving the user's Quality of Experience (QoE). In this paper, we experiment with ABR algorithms for VR streaming using Air Light VR (ALVR), an open-source VR streaming solution. We extend ALVR with a comprehensive set of metrics that provide a robust characterization of the network's state, enabling more informed bitrate adjustments. To demonstrate the utility of these performance indicators, we develop and test the Network-aware Step-wise ABR algorithm for VR streaming (NeSt-VR). Results validate the accuracy of the newly implemented network performance metrics and demonstrate NeSt-VR's video bitrate adaptation capabilities.

Open-source Large Language Models enable projects such as NASA SciX (i.e., NASA ADS) to think out of the box and try alternative approaches for information retrieval and data augmentation, while respecting data copyright and users' privacy. However, when large language models are directly prompted with questions without any context, they are prone to hallucination. At NASA SciX we have developed an experiment where we created semantic vectors for our large collection of abstracts and full-text content, and we designed a prompt system to ask questions using contextual chunks from our system. Based on a non-systematic human evaluation, the experiment shows a lower degree of hallucination and better responses when using Retrieval Augmented Generation. Further exploration is required to design new features and data augmentation processes at NASA SciX that leverages this technology while respecting the high level of trust and quality that the project holds.

Simple, portable and low-cost experiments as RC and RL series circuits are proposed to experiment with DC circuits. Very common elements are used: a few electronics components (resistors, capacitors, coils and connecting wires) and two smartphones. We consider the charging and discharging of a capacitor in the RC circuit and also that of coil in the RL circuit. Using a smartphone as an oscilloscope we observe voltages variations which are the transient response to a square signal generated in the second smartphone. These voltage variations are directly related to the electrostatic or magnetic energy stored in the circuits. The experimental data have been collected with the smartphone used as an oscilloscope and corroborated with theoretical predictions based on Kirchhoff's laws. The comparison showed differences of the order of the 1\% or less between the calculated capacitance or inductance compared to the manufacturer values. This approach which avoids the use of expensive signal generators, oscilloscopes, or any specialized hardware can be performed in less-favored contexts and even as a home assignment.

We revisit, implement, and experiment with a beautiful algorithm, due to Calabi and Wilf for the random generation of subspaces over a finie field.

Large Language Models (LLM) have become sophisticated enough that complex computer programs can be created through interpretation of plain English sentences and implemented in a variety of modern languages such as Python, Java Script, C++ and Spreadsheets. These tools are powerful and relatively accurate and therefore provide broad access to computer programming regardless of the background or knowledge of the individual using them. This paper presents a series of experiments with ChatGPT to explore the tool's ability to produce valid spreadsheet formulae and related computational outputs in situations where ChatGPT has to deduce, infer and problem solve the answer. The results show that in certain circumstances, ChatGPT can produce correct spreadsheet formulae with correct reasoning, deduction and inference. However, when information is limited, uncertain or the problem is too complex, the accuracy of ChatGPT breaks down as does its ability to reason, infer and deduce. This can also result in false statements and "hallucinations" that all subvert the process of creating spreadsheet formulae.

We study randomized experiments in a service system when stochastic congestion can arise from temporarily limited supply or excess demand. Such congestion gives rise to cross-unit interference between the waiting customers, and analytic strategies that do not account for this interference may be biased. In current practice, one of the most widely used ways to address stochastic congestion is to use switchback experiments that alternatively turn a target intervention on and off for the whole system. We find, however, that under a queueing model for stochastic congestion, the standard way of analyzing switchbacks is inefficient, and that estimators that leverage the queueing model can be materially more accurate. Additionally, we show how the queueing model enables estimation of total policy gradients from unit-level randomized experiments, thus giving practitioners an alternative experimental approach they can use without needing to pre-commit to a fixed switchback length before data collection.

Decentralised Machine Learning (DML) enables collaborative machine learning without centralised input data. Federated Learning (FL) and Edge Inference are examples of DML. While tools for DML (especially FL) are starting to flourish, many are not flexible and portable enough to experiment with novel processors (e.g., RISC-V), non-fully connected network topologies, and asynchronous collaboration schemes. We overcome these limitations via a domain-specific language allowing us to map DML schemes to an underlying middleware, i.e. the FastFlow parallel programming library. We experiment with it by generating different working DML schemes on x86-64 and ARM platforms and an emerging RISC-V one. We characterise the performance and energy efficiency of the presented schemes and systems. As a byproduct, we introduce a RISC-V porting of the PyTorch framework, the first publicly available to our knowledge.

Teaching digital manufacturing at scale using MOOCs has opened opportunities for IMT, a network of French graduate engineering schools, to work closely with a community of learners and educators in physical spaces called Fab Labs. By setting up a cohort of lifelong learning trainees taking the MOOC online and attending hands-on in-person workshops, IMT to experiment blended learning models and hybrid skills certification for project-based STEM courses.

This paper summarizes the results of experimenting with Universal Dependencies (UD) adaptation of an Unsupervised, Compositional and Recursive (UCR) rule-based approach for Sentiment Analysis (SA) submitted to the Shared Task at Rest-Mex 2023 (Team Olga/LyS-SALSA) (within the IberLEF 2023 conference). By using basic syntactic rules such as rules of modification and negation applied on words from sentiment dictionaries, our approach exploits some advantages of an unsupervised method for SA: (1) interpretability and explainability of SA, (2) robustness across datasets, languages and domains and (3) usability by non-experts in NLP. We compare our approach with other unsupervised approaches of SA that in contrast to our UCR rule-based approach use simple heuristic rules to deal with negation and modification. Our results show a considerable improvement over these approaches. We discuss future improvements of our results by using modality features as another shifting rule of polarity and word disambiguation techniques to identify the right sentiment words.

Introduction Data imbalance is one of the crucial issues in big data analysis with fewer labels. For example, in real-world healthcare data, spam detection labels, and financial fraud detection datasets. Many data balance methods were introduced to improve machine learning algorithms' performance. Research claims SMOTE and SMOTE-based data-augmentation (generate new data points) methods could improve algorithm performance. However, we found in many online tutorials, the valuation methods were applied based on synthesized datasets that introduced bias into the evaluation, and the performance got a false improvement. In this study, we proposed, a new evaluation framework for imbalanced data learning methods. We have experimented on five data balance methods and whether the performance of algorithms will improve or not. Methods We collected 8 imbalanced healthcare datasets with different imbalanced rates from different domains. Applied 6 data augmentation methods with 11 machine learning methods testing if the data augmentation will help with improving machine learning performance. We compared the traditional data augmentation evaluation methods with our proposed cross-validation eval