Newest and Lowest Rated Papers
Advanced filtersTopic: AllExperimental
Topic
Topics are auto-detected from title, abstract, and metadata and may be imperfect.
Literary Hub » All or Nothing: Deborah Levy on Marguerite Duras’s The Lover
October 7, 2024
Literary Hub·2024
No abstract available
★ 3.0 (1)
View paper →Large Language Model Prediction Capabilities: Evidence from a Real-World Forecasting Tournament
Philipp Schoenegger, Peter S. Park
arXiv·2023
Accurately predicting the future would be an important milestone in the capabilities of artificial intelligence. However, research on the ability of large language models to provide probabilistic predictions about future events remains nascent. To empirically test this ability, we enrolled OpenAI's state-of-the-art large language model, GPT-4, in a three-month forecasting tournament hosted on the Metaculus platform. The tournament, running from July to October 2023, attracted 843 participants and covered diverse topics including Big Tech, U.S. politics, viral outbreaks, and the Ukraine conflict. Focusing on binary forecasts, we show that GPT-4's probabilistic forecasts are significantly less accurate than the median human-crowd forecasts. We find that GPT-4's forecasts did not significantly differ from the no-information forecasting strategy of assigning a 50% probability to every question. We explore a potential explanation, that GPT-4 might be predisposed to predict probabilities close to the midpoint of the scale, but our data do not support this hypothesis. Overall, we find that GPT-4 significantly underperforms in real-world predictive tasks compared to median human-crowd forecasts. A potential explanation for this underperformance is that in real-world forecasting tournaments, the true answers are genuinely unknown at the time of prediction; unlike in other benchmark tasks like professional exams or time series forecasting, where strong performance may at least partly be due to the answers being memorized from the training data. This makes real-world forecasting tournaments an ideal environment for testing the generalized reasoning and prediction capabilities of artificial intelligence going forward.
★ 3.0 (1)
View paper →Oxidation products of polyunsaturated fatty acids in infant formulas compared to human milk – A preliminary study
Marie‐Caroline Michalski, Catherine Calzada, Asami Makino, Sabine Michaud, Michel Guichardant
Molecular Nutrition & Food Research·2008·69 citations
<jats:title>Abstract</jats:title><jats:p>Information about lipid oxidation in fresh and stored human milk compared with infant formulas is scarce. We aimed to assess<jats:italic> n</jats:italic>‐6 and<jats:italic> n</jats:italic>‐3 PUFA oxidation in these milks by measuring the 4‐hydroxynonenal (4‐HNE) and 4‐hydroxyhexenal (4‐HHE) content. Human milk samples (<jats:italic>n </jats:italic>= 4), obtained from volunteer mothers, were analyzed fresh and after 1 wk at 4°C or 24 h at 18°C. Vitamin E and malondialdehyde (MDA) were measured by HPLC and fatty acid profile by GC. The 4‐HHE and 4‐HNE contents were measured by GC‐MS. Infant formulas (<jats:italic>n</jats:italic> = 10) were tested; their fat droplet size was measured by laser light scattering and observed by confocal laser scanning microscopy. Human milk samples contained 31.0 ± 6.3 g/L of lipids and 1.14 ± 0.26 mg/L of vitamin E. Fat droplets were smaller in infant formulas than reported in human milk. The (4‐HHE/<jats:italic>n</jats:italic>‐3 PUFA) ratio was 0.19 ± 0.01 μg/g in fresh human milk (unchanged after storage)<jats:italic> versus</jats:italic> 3.6 ± 3.1 μg/g in dissolved powder formulas and 4.3 ± 3.8 μg/g in liquid formula. (4‐HNE/<jats:italic>n</jats:italic>‐6 PUFA) was 0.004 ± 0.000 μg/g in fresh milk (0.03 ± 0.01 μg/g after storage) <jats:italic>versus</jats:italic> 1.1 ± 1.0 μg/g in dissolved powder formulas and 0.2 ± 0.3 μg/g in liquid formula. Infant formulas also contained more MDA than human milk.<jats:italic> n</jats:italic>‐3 PUFA were more prone to oxidation than <jats:italic>n</jats:italic>‐6 PUFA. Whether threshold levels of 4‐HHE and 4‐HNE would be of health concern should be elucidated.</jats:p>
★ 3.0 (1)
View paper →