Related papers: Dependence of technological improvement on artifac…
Human data annotation, especially when involving experts, is often treated as an objective reference. However, many annotation tasks are inherently subjective, and annotators' judgments may evolve over time. This study investigates changes…
How should researchers analyze randomized experiments in which the main outcome is latent and measured in multiple ways but each measure contains some degree of error? We first identify a critical study-specific noncomparability problem in…
As AI increasingly enters the classroom, what changes when students collaborate with algorithms instead of peers? We analyzed 36 undergraduate students learning graph theory through peer collaboration (n=24) or AI assistance (n=12), using…
Technology is essential to innovation and economic prosperity. Understanding technological changes can guide innovators to find new directions of design innovation and thus make breakthroughs. In this work, we construct a technology fitness…
A growing literature has examined whether innovation is becoming less disruptive, spanning diverse domains and data sources and using a range of methodologies. This paper provides an inventory of 105 studies exploring this question. The…
The foremost challenge to causal inference with real-world data is to handle the imbalance in the covariates with respect to different treatment options, caused by treatment selection bias. To address this issue, recent literature has…
Technological knowledge evolves not only through the generation of new ideas, but also through the reinterpretation of existing ones. Reinterpretations lead to changes in the classification of knowledge, that is, reclassification. This…
High-dimensional measurements are often correlated which motivates their approximation by factor models. This holds also true when features are engineered via low-dimensional interactions or kernel tricks. This often results in over…
Considering collaborative patent development, we provide micro-level evidence for innovation through exchanges of differentiated knowledge. Knowledge embodied in a patent is proxied by word pairs appearing in its abstract, while novelty is…
Species interactions (ranging from direct predator prey relationships to indirect effects mediated by the environment) are central to ecosystem balance and biodiversity. While empirical methods for measuring these interactions exist, their…
Method-level comments are critical for improving code comprehension and supporting software maintenance. With advancements in large language models (LLMs), automated comment generation has become a major research focus. However, existing…
The prevailing assumption in agent orchestration is that more context is better. We test this on multi-agent software design across 10 tasks, 7 context-injection conditions, and over 2,700 runs, and find a crossover effect: the same…
Dynamical systems across many disciplines are modeled as interacting particles or agents, with interaction rules that depend on a very small number of variables (e.g. pairwise distances, pairwise differences of phases, etc...), functions of…
Agentic AI workflows (systems that autonomously plan and act) are becoming widespread, yet their task success rate on complex tasks remains low. A promising solution is inference-time alignment, which uses extra compute at test time to…
Generative language models are promising for assisting human writing in various domains. This manuscript aims to build generative language models in the patent domain and evaluate model performance from a human-centric perspective. The…
Correspondence identifies relationships among objects via similarities among their components; it is ubiquitous in the analysis of spatial datasets, including images, weather maps, and computational simulations. This paper develops a novel…
In language model training, it is desirable to equip models with capabilities from various tasks. However, it is not clear how to directly obtain the right data mixtures for these capabilities as the relationship between data and tasks is…
Identifying changes in contact during contact-rich manipulation can detect task state or errors, enabling improved robustness and autonomy. The ability to detect contact is affected by the mechatronic design of the robot, especially its…
We study an interacting particle system whose dynamics depends on an interacting random environment. As the number of particles grows large, the transition rate of the particles slows down (perhaps because they share a common resource of…
We study long-run progress in artificial intelligence in a quantitative way. Many measures, including traditional ones such as patents and publications, machine learning benchmarks, and a new Aggregate State of the Art in ML (or ASOTA)…