Claims Browser

Most current traffic models fix in advance which behavioral variables interact and how, absorbing behavior outside that form as noise, while flexible models tend to lose interpretability.

"Most models, however, fix in advance which behavioral variables interact and how. Behavior outside that form is absorbed as noise, while models flexible enough to capture it tend to lose interpretability."

A quantum-inspired representation of driver behavior can combine properties usually treated separately: continuous, probabilistic, context-dependent, history-dependent representation with learned interactions among behavioral variables.

"We introduce a quantum-inspired representation of driver behavior that combines properties usually treated separately or in part: it is continuous, probabilistic, context-dependent, history-dependent, and represents interactions among behavioral variables as learned from data."

When trained without supervision on the I-24 MOTION dataset, the quantum-inspired framework recovers three interpretable driving profiles representing free flow, transition, and congestion regimes.

"Trained without supervision on the I-24 MOTION dataset, the framework recovers three interpretable driving profiles representing three regimes: free flow, transition, and congestion."

The quantum-inspired driver representation reproduces known macroscopic traffic phenomena, aligning with the fundamental diagram and reproducing hysteresis loops.

"The same representation also reproduces known macroscopic phenomena, aligning with the fundamental diagram and reproducing hysteresis loops."

A matching lower bound of Ω(d log(1+T/d)) shows that the Ω(log T) factor is unavoidable for general OCO

"We also prove a matching lower bound of $Ω(d\log (1+T/d))$, showing that the $Ω(\log T)$ factor is unavoidable."

For general OCO over a d-dimensional compact convex set, an algorithm achieves O(d log(1+T/d)) alternating regret, improving the previous best of O-tilde(d^(2/3) T^(1/3))

"For general OCO over a $d$-dimensional compact convex set, we give an algorithm with $O(d\log (1+T/d))$ alternating regret, improving the previous best of $\widetilde{O}(d^{2/3}T^{1/3})$."

This is the first uncoupled learning dynamics with O(1/T) convergence to CCE in two-player general-sum games, while all prior works suffer additional log T factors

"This is the first uncoupled learning dynamics with $O(1/T)$ convergence to CCE in two-player general-sum games, while all prior works suffer additional $\log T$ factors."

Alternating learning dynamics achieve O(log d / T) convergence to Nash equilibria in two-player zero-sum games and O(log d / T) convergence to coarse correlated equilibria in two-player general-sum games

"As a result, we obtain alternating learning dynamics with $O(\log d /T)$ convergence to Nash equilibria in two-player zero-sum games and $O(\log d /T)$ convergence to coarse correlated equilibria in two-player general-sum games."

The constant regret bound significantly improves previous results which had O(log^(2/3) d · T^(1/3)) regret

"Our constant regret bound significantly improves previous results with $O(\log ^{2/3}d \cdot T^{1/3})$ regret [Cevher, Cutkosky, Kavis, Piliouras, Skoulakis, Viano, NeurIPS 2023, Hait, Li, Luo, Zhang, COLT 2025]."

An algorithm achieves O(log d) alternating regret for online linear optimization over the probability simplex, which remains constant for any time horizon T

"For OLO over the probability simplex $Δ_d$, we give an algorithm with $O(\log d)$ alternating regret that remains a constant for any time horizon $T$, and a matching lower bound."

The dual PCN approach results in more stable, continuous, online learning for time series anomaly detection

"our approach results in more stable, continuous, online learning"

Training a Generative PCN paired with an Encoding PCN in parallel to match neural activations enables training without sequential propagation

"we introduce a training technique that pairs a Generative PCN with a support Encoding PCN. The two PCNs are trained in parallel to match their neural activations, without sequential propagation"

Sequential backwards error propagation is the main bottleneck of Predictive Coding Networks

"the main bottleneck of PC Networks (PCN) is the sequential backwards error propagation"

Predictive Coding Networks enable parallelizable neural network layer updates

"Predictive Coding (PC) is a neural learning paradigm that enables parallelizable neural network layer updates"

Microsoft's MAI-Thinking-1, a 35-billion-active-parameter reasoning model trained from scratch on commercially licensed data, matches Claude Opus 4.6 on SWE Bench Pro

"Leading them was MAI-Thinking-1, a 35-billion-active-parameter reasoning model Microsoft says it trained from scratch on commercially licensed data and that matches Claude Opus 4.6 on SWE Bench Pro."

Mythos 5's protein design experts sped up aspects of drug design by around 10 times

"its protein design experts sped up aspects of drug design by around 10 times."

Fable 5 migrated a 50-million-line Ruby codebase in a day, work that would have taken a team over two months by hand

"Stripe reported in the launch materials that Fable 5 migrated a 50-million-line Ruby codebase in a day, work that would have taken a team over two months by hand."

Anthropic released Claude Fable 5 on June 9, a Mythos-class model made safe for general use, alongside Claude Mythos 5 with lifted safeguards

"Anthropic moved first, releasing Claude Fable 5 on June 9, a Mythos-class model made safe for general use, alongside Claude Mythos 5 — the same model with safeguards lifted in some areas."