eamag/beyond_benchmarks · Datasets at Hugging Face

these findings corroborate that contrastive representation learning benefits not from accurate mi

estimation

density-ratio estimation

these findings corroborate that contrastive representation learning benefits not from accurate mi estimation per se but from the learning of structured density ratios

it can also be approximately scaled to the global agn accretion rate as dot m _ mathrm vis propto dot

m

accretion rate

it can also be approximately scaled to the global agn accretion rate as dot m _ mathrm vis propto dot m _1 with the scaling coefficients in both forms determined by the specific dynamical configuration

while optical flow a computer vision technique for estimating pixel wise

motion

point tracking

while optical flow a computer vision technique for estimating pixel wise motion between consecutive images has advanced rapidly in computer vision its applicability to geophysical problems and to satellite sar imagery remains underexplored

these data provide important constraints for chemodynamical models of massive

protostellar

stellar mass

these data provide important constraints for chemodynamical models of massive protostellar cores

first an equivalence between klyachko-can-binicio u g lu-shumovsky kcbs vectors and the pyramid upb is shown and then by constructing a one parameter family of upb vectors a quantitative connection between contextuality strength and bound entanglement of

states

quantum advantage

first an equivalence between klyachko-can-binicio u g lu-shumovsky kcbs vectors and the pyramid upb is shown and then by constructing a one parameter family of upb vectors a quantitative connection between contextuality strength and bound entanglement of states associated with the corresponding upb is demonstrated

lastly we do not observe a significant curse of multilinguality as the number of training

languages

vision-language models

lastly we do not observe a significant curse of multilinguality as the number of training languages increases in models at this scale

sensitivity analysis for treatment effects in

difference-in-differences

average treatment effect

sensitivity analysis for treatment effects in difference-in-differences models using riesz representation

however traditional physics-based simulation

methods

classical simulation

however traditional physics-based simulation methods for modeling these processes often suffer from prohibitive computational costs

we present high-resolution x-ray spectroscopy of the merging galaxy

cluster

star clusters

we present high-resolution x-ray spectroscopy of the merging galaxy cluster abell 3667 with textit xrism resolve

distributed quantum computing dqc provides a promising route toward scalable quantum

computation

quantum key distribution

distributed quantum computing dqc provides a promising route toward scalable quantum computation where entanglement-assisted locc and circuit knitting represent two complementary approaches

prejudice driven spite a discontinuous phase

transition

game theory

prejudice driven spite a discontinuous phase transition in ultimatum game

this survey provides an analysis of current methodologies integrating legal and logical specifications into the perception prediction and planning modules of automated

driving

autonomous driving

this survey provides an analysis of current methodologies integrating legal and logical specifications into the perception prediction and planning modules of automated driving systems

reconstructing evolutionary histories and estimating the rate of evolution from molecular sequence data is of central importance in

evolutionary

phylogenetic tree

reconstructing evolutionary histories and estimating the rate of evolution from molecular sequence data is of central importance in evolutionary biology and infectious disease research

we study bilevel optimization problems where the lower-level problems are strongly

convex

strongly convex

we study bilevel optimization problems where the lower-level problems are strongly convex and have coupled linear constraints

we show an algorithm based on a trust-region method with an oracle that can be implemented using widetilde o d 1 3 linear system solves improving the widetilde o n 1 3

oracle

-approximation algorithm

we show an algorithm based on a trust-region method with an oracle that can be implemented using widetilde o d 1 3 linear system solves improving the widetilde o n 1 3 oracle by

our results confirm the long-standing intuition that posterior inconsistency in

density

density estimation

our results confirm the long-standing intuition that posterior inconsistency in density estimation is not a natural phenomenon but rather an artifact of pathological prior constructions

reinforcement learning rl fine-tuning of large language models llms often suffers from instability due to the numerical mismatch between the

training

reinforcement learning rl

reinforcement learning rl fine-tuning of large language models llms often suffers from instability due to the numerical mismatch between the training and inference policies

model-free filtering in high dimensions via

projection

diffusion models

model-free filtering in high dimensions via projection and score-based diffusions

using quantum mechanical density functional theory calculations we propose a prototypical bilayer heterostructure composed of hexagonal boron nitride hbn and silicon carbide sic characterized by a

lattice

quantum materials

using quantum mechanical density functional theory calculations we propose a prototypical bilayer heterostructure composed of hexagonal boron nitride hbn and silicon carbide sic characterized by a lattice mismatch of 18

in this two-paper series we present a straightforward mathematical model for synthesizing quasar absorption line profiles from sight lines through idealized spatial-kinematic models of the

circumgalactic

galaxy cgm

in this two-paper series we present a straightforward mathematical model for synthesizing quasar absorption line profiles from sight lines through idealized spatial-kinematic models of the circumgalactic medium cgm and their host galaxies

this paper develops a unified framework for identifying

spatial

treatment effect boundaries

this paper develops a unified framework for identifying spatial and temporal boundaries of treatment effects in difference-in-differences designs

to capture the long-term dependency and complex dynamics of

eeg

brain decoding

to capture the long-term dependency and complex dynamics of eeg we propose a hybrid encoder combining a mamba-like linear attention channel encoder and a spatiotemporal dynamics model

the phase transition is quite robust and becomes progressively conspicuous in the limit of large population size where deterministic evolutionary

game

game theory

the phase transition is quite robust and becomes progressively conspicuous in the limit of large population size where deterministic evolutionary game dynamics viz

however the up-to-date runtime estimates for utility-scale applications on certain quantum

hardware

quantum computing

however the up-to-date runtime estimates for utility-scale applications on certain quantum hardware systems are in the order of years rendering quantum computations impractical

our empirical analysis reveals that covid-19 elevated

network

network fragility

our empirical analysis reveals that covid-19 elevated network fragility measured by the algebraic connectivity lambda_2 of the system laplacian by 26

communication impact is quantied by a capacity-motivated lower bound obtained from the linear minimum mean-squared error error covariance with a mismatched

channel

channel state information

communication impact is quantied by a capacity-motivated lower bound obtained from the linear minimum mean-squared error error covariance with a mismatched channel estimate

this paper investigates the construction of channel knowledge map ckm from sparse

channel

channel state information csi

this paper investigates the construction of channel knowledge map ckm from sparse channel measurements

on the theoretical front we observe that studying the approximate rank of

language

large language

on the theoretical front we observe that studying the approximate rank of language models in the sense discussed above yields a simple universal abstraction whose theoretical predictions parallel our experiments

quantum computation with d -level quantum systems also known as qudits benefits from the possibility to use a richer

computational

quantum batteries

quantum computation with d -level quantum systems also known as qudits benefits from the possibility to use a richer computational space compared to qubits

extensions to longitudinal data dynamic treatment regimes and multiplicative instrumental

variables

treatment effect

extensions to longitudinal data dynamic treatment regimes and multiplicative instrumental variables are further developed

enhancing the spatial awareness of underwater

vehicles

autonomous driving

enhancing the spatial awareness of underwater vehicles is key to reducing piloting risks and enabling greater autonomy

in this model hassidim and singer 2017 design a 1-1 e -approximation algorithm for monotone submodular maximization subject to a cardinality constraint and huang et al 2022 design a 1-1 e 2 -approximation algorithm for monotone

submodular

monotone submodular

in this model hassidim and singer 2017 design a 1-1 e -approximation algorithm for monotone submodular maximization subject to a cardinality constraint and huang et al 2022 design a 1-1 e 2 -approximation algorithm for monotone submodular maximization subject to any arbitrary matroid constraint

the training set is strategically constructed to promote the descriptor s generalizability which is systematically evaluated by verification and validation during the

training

machine learning

the training set is strategically constructed to promote the descriptor s generalizability which is systematically evaluated by verification and validation during the training process

instrumental variable methods are fundamental to

causal

causal effects

instrumental variable methods are fundamental to causal inference when treatment assignment is confounded by unobserved variables

the review is structured to first establish the theoretical foundation for analyzing these complex systems examining both structural models of complex networks and physical models of social

dynamics

opinion dynamics

the review is structured to first establish the theoretical foundation for analyzing these complex systems examining both structural models of complex networks and physical models of social dynamics e

we demonstrate how the metallicity dependence of the yields can be mathematically considered as a system-dependent delay time approximately equal to the system s depletion

time

star-forming region

we demonstrate how the metallicity dependence of the yields can be mathematically considered as a system-dependent delay time approximately equal to the system s depletion time that when combined with system-independent delay times arising from stellar evolutionary channels produces the separation of different systems based on their star formation efficiency and mass-loading factor

recap reproducing copyrighted data from llms training with an

agentic

llm agents

recap reproducing copyrighted data from llms training with an agentic pipeline

here we compare three different ar navigation aids on-screen compass on-screen radar and in-world vertical arrows in a wide-area outdoor user study n 24 where participants search for hidden virtual target items amongst physical and

virtual

virtual reality

here we compare three different ar navigation aids on-screen compass on-screen radar and in-world vertical arrows in a wide-area outdoor user study n 24 where participants search for hidden virtual target items amongst physical and virtual objects

3 as a result of the cancellation effect of randomly-oriented magnetic fields

induced

magnetic field

3 as a result of the cancellation effect of randomly-oriented magnetic fields induced by sloshing-driven turbulence

we present the first large scale benchmark of 48 deep learning optical

flow

optical flow

we present the first large scale benchmark of 48 deep learning optical flow models on radarsat 2 scansar sea ice imagery evaluated with endpoint error epe and fl all metrics against gnss tracked buoys

the tree supports text polylog n -time operations and requires a static lookup table of size text poly n text polylog u -- but in exchange for these the

tree

tree embedding

the tree supports text polylog n -time operations and requires a static lookup table of size text poly n text polylog u -- but in exchange for these the tree is able to achieve a remarkable space guarantee

on the other hand we study nonparametric analogues of this problem in smooth regression and

density

diffusion models

on the other hand we study nonparametric analogues of this problem in smooth regression and density models

to standardize this study we curate the evaluation data into mme-cof a compact

benchmark

evaluation metrics

to standardize this study we curate the evaluation data into mme-cof a compact benchmark that enables in-depth and thorough assessment of chain-of-frame cof reasoning

these systems exhibit high bulge-to-total b t

light

light curves

these systems exhibit high bulge-to-total b t light ratios 0

this paper presents a pipeline integrating fine-tuned large language models llms with named

entity

natural language processing

this paper presents a pipeline integrating fine-tuned large language models llms with named entity recognition ner for efficient domain-specific text summarization and tagging

in this study we used a data-driven network approach to examine whether resting-state

eeg

fmri data

in this study we used a data-driven network approach to examine whether resting-state eeg connectivity patterns differentiate individuals according to their creative abilities

experimental results show that in environments with minimal vertical

variation

higher-order visual

experimental results show that in environments with minimal vertical variation the proposed 3d model matches the performance of a 2d baseline yet as 3d complexity increases it yields substantially more distinct place fields and markedly reduces spatial aliasing

however causal inference from observational data relies on untestable

assumptions

causal inference

however causal inference from observational data relies on untestable assumptions about the data-generating process such as the absence of unobserved confounders

to optimize this trade-off we pose a non-convex

joint

optimal power flow

to optimize this trade-off we pose a non-convex joint ota power-control design and develop an efficient successive convex approximation sca algorithm that requires only statistical csi at the base station

keywords data visualization radar charts combinatorial optimization minimax optimization

feature

radar charts

keywords data visualization radar charts combinatorial optimization minimax optimization feature ordering

across quantitative metrics and qualitative assessments our method achieves superior generalization and improved performance relative to pose- or trajectory-conditioned baselines advancing practical 4d

world

world models

across quantitative metrics and qualitative assessments our method achieves superior generalization and improved performance relative to pose- or trajectory-conditioned baselines advancing practical 4d world modeling from casual videos

numerical comparisons show that our approach meets the stringent timeliness requirement and achieves a six-fold reduction in rb utilization or a 6 db energy saving

compared

computationally efficient

numerical comparisons show that our approach meets the stringent timeliness requirement and achieves a six-fold reduction in rb utilization or a 6 db energy saving compared to benchmarks

our findings identify magnetic tidal coupling as a novel strong-gravity effect and establish its importance for the resonant dynamics of

compact-object

tidal disruption

our findings identify magnetic tidal coupling as a novel strong-gravity effect and establish its importance for the resonant dynamics of compact-object binaries near smbhs

witnessing genuine multipartite entanglement in phase

space

entanglement entropy

witnessing genuine multipartite entanglement in phase space with controlled gaussian unitaries

distributed quantum computing dqc provides a promising route toward scalable quantum

computation

quantum error correction

distributed quantum computing dqc provides a promising route toward scalable quantum computation where entanglement-assisted locc and circuit knitting represent two complementary approaches

artificial intelligence has advanced significantly through deep learning reinforcement learning and large

language

large language

artificial intelligence has advanced significantly through deep learning reinforcement learning and large language and vision models

instead of extracting fingerprints directly from raw csi measurements csi2q first transforms frequency-domain

csi

csi dataset

instead of extracting fingerprints directly from raw csi measurements csi2q first transforms frequency-domain csi measurements into time-domain signals that share the same feature space with iq samples

to overcome the challenge this work suggests leveraging a multi-frequency

neural

neural network

to overcome the challenge this work suggests leveraging a multi-frequency neural network named mfnn embedding prior physical knowledge into the network

we validate this framework across multiple lattice sizes showing that it preserves physical fidelity while outperforming monte

carlo

monte carlo

we validate this framework across multiple lattice sizes showing that it preserves physical fidelity while outperforming monte carlo generation in speed as the lattice size increases

this approach is particularly relevant for predicting the emergence of paradoxical neural representations such as discordant visual illusions that occur in response to overt

sensory

predictive processing

this approach is particularly relevant for predicting the emergence of paradoxical neural representations such as discordant visual illusions that occur in response to overt sensory stimuli

the training process of a deep neural network is partitioned between devices and the ap where a dt replica is

activated

deep reinforcement learning

the training process of a deep neural network is partitioned between devices and the ap where a dt replica is activated to replace uds with insufficient local computation capabilities

to formalize this we extend the concept of data informativity by requiring the existence of a controller that stabilizes all systems consistent with the data and the

prior

data-driven stabilization

to formalize this we extend the concept of data informativity by requiring the existence of a controller that stabilizes all systems consistent with the data and the prior knowledge

these include the feeling of spatial extendedness temporal flow of

objects

higher-order visual

these include the feeling of spatial extendedness temporal flow of objects binding general concepts with particular configurations of features and of qualia such as colors and sounds

more importantly we find that 12 of these mechanisms can be unambiguously identified by deriving a scaling law that expresses the

third-order

third-order nonlinear

more importantly we find that 12 of these mechanisms can be unambiguously identified by deriving a scaling law that expresses the third-order nonlinear hall conductivity as a polynomial in the linear longitudinal conductivity

spiral features appear across stellar populations of different ages confirming their density-wave nature and producing coherent spirals in both

metallicity

star-forming region

spiral features appear across stellar populations of different ages confirming their density-wave nature and producing coherent spirals in both metallicity and mean stellar age distributions-consistent with recent gaia observations of the milky way

large language models llms are widely used in generative

applications

models llms

large language models llms are widely used in generative applications such as chatting code generation and reasoning

in this paper we study linear control systems with positive

bounded

control systems

in this paper we study linear control systems with positive bounded orbits

we demonstrate that translation difficulty correlates with physical coupling achieving near-perfect fidelity for mappings from gas to

dark

dark matter

we demonstrate that translation difficulty correlates with physical coupling achieving near-perfect fidelity for mappings from gas to dark matter and mappings involving astro-chemical components such as total gas to hi content while identifying fundamental challenges in weakly constrained tasks such as gas to stellar mass mappings

in many practical settings however user interactions are limited or costly making offline

preference

preference optimization

in many practical settings however user interactions are limited or costly making offline preference learning necessary

in this paper we extend such a theory to the case of infinite horizon optimal

control

linear control

in this paper we extend such a theory to the case of infinite horizon optimal control problems which are very common in particular in economic applications

in this paper we propose adasdbo a fully problem-parameter-free algorithm for decentralized

bilevel

bilevel optimization

in this paper we propose adasdbo a fully problem-parameter-free algorithm for decentralized bilevel optimization with a single-loop structure

our results lay a fundamental understanding of how pre-trained llms manipulate numbers and outline the

potential

llm raters

our results lay a fundamental understanding of how pre-trained llms manipulate numbers and outline the potential of more accurate probing techniques in addressed refinements of llms architectures

temporal disease dynamics and treatment switch effects are then captured through a

mixed-effects

mixed-effects regression

temporal disease dynamics and treatment switch effects are then captured through a mixed-effects regression model applied to latent representations

however many existing math competitions are becoming less effective for assessing top-tier llms due to performance

saturation

superior performance

however many existing math competitions are becoming less effective for assessing top-tier llms due to performance saturation e

in more complex scenarios we show that thinking of

selection

evolutionary dynamics

in more complex scenarios we show that thinking of selection over discrete generations has significant advantages

these findings suggest that geometry and tuning encode brain-region- or model-family-specific signatures while linearly decodable information tends to be more globally shared across

regions

functional connectivity

these findings suggest that geometry and tuning encode brain-region- or model-family-specific signatures while linearly decodable information tends to be more globally shared across regions or models

this is an exponential improvement over previous results and only a

polylogarithmic

mathrm polylog

this is an exponential improvement over previous results and only a polylogarithmic factor away from the lower bound

from amateur to master infusing knowledge into

llms

models llms

from amateur to master infusing knowledge into llms via automated curriculum learning

this paper investigates large-population stochastic

control

linear control

this paper investigates large-population stochastic control problems in which agents share their state information and cooperate to minimize a convex cost functional

overall they are not yet reliable as standalone zero-shot reasoners but exhibit encouraging signs as complementary visual engines alongside dedicated

reasoning

reasoning curriculum

overall they are not yet reliable as standalone zero-shot reasoners but exhibit encouraging signs as complementary visual engines alongside dedicated reasoning models

these findings support a clear takeaway improving representation learning is a direct and useful path to robust

world

world models

these findings support a clear takeaway improving representation learning is a direct and useful path to robust world models delivering reliable long-horizon predictions without enlarging the dynamics module

more importantly the thinking structure in this protocol can be further optimized through

reinforcement

reasoning capabilities

more importantly the thinking structure in this protocol can be further optimized through reinforcement learning

the introduction of integrated sensing and communications

isac

communication isac

the introduction of integrated sensing and communications isac in cellular systems is not expected to result in a shift away from the popular choice of cost- and energy-efficient analog or hybrid beamforming structures

this paper investigates whether pretrained language models including large language models possess similar

capabilities

language models

this paper investigates whether pretrained language models including large language models possess similar capabilities for loanword identification

in this study we propose a mosquito-borne disease model incorporating odor-baited traps to examine the impact of odor on

disease

disease transmission

in this study we propose a mosquito-borne disease model incorporating odor-baited traps to examine the impact of odor on disease transmission

specifically when f x acts safely but u0 acts unsafely the gain can be decreased by up to half and when f x acts unsafely we establish that if u0 acts safely the gain can be increased arbitrarily whereas if u0 acts unsafely the control recovers the full

gain

control strategy

specifically when f x acts safely but u0 acts unsafely the gain can be decreased by up to half and when f x acts unsafely we establish that if u0 acts safely the gain can be increased arbitrarily whereas if u0 acts unsafely the control recovers the full gain margin 1 2 inf

we train neural networks to compress perceptual and semantic factors of stimuli measuring lossiness using the

mathematical

deep learning

we train neural networks to compress perceptual and semantic factors of stimuli measuring lossiness using the mathematical framework underlying compression

this work presents contributions to a state-of-the-art

stability

data-driven stabilization

this work presents contributions to a state-of-the-art stability verification procedure for nnc-controlled systems which relies on semialgebraic-set-based input-output modeling to pose the search for a lyapunov function as an optimization problem

we formalize the separation of perception and

decision

debiased machine learning

we formalize the separation of perception and decision define perceptual properties independent of objectives or reparameterizations and prove that pel updates preserving sufficient invariants are orthogonal to bayes task-risk gradients

identifying asymptomatic nodes in network

epidemics

viral replication

identifying asymptomatic nodes in network epidemics using graph neural networks

this paper develops a unified theoretical framework for detecting and

estimating

average treatment effect

this paper develops a unified theoretical framework for detecting and estimating boundaries in treatment effects across both spatial and temporal dimensions

statistical physics of deep learning optimal learning of a

multi-layer

neural network

statistical physics of deep learning optimal learning of a multi-layer perceptron near interpolation

a sliding-window filter for online continuous-time continuum

robot

robotic systems

a sliding-window filter for online continuous-time continuum robot state estimation

beyond reality designing personal experiences and interactive narratives in

ar

augmented reality

beyond reality designing personal experiences and interactive narratives in ar theater

we have investigated the orbital-dependent

electronic

electronic structure

we have investigated the orbital-dependent electronic states of cd6ce a prototype of strongly correlated rare-earth-based tsai-type quasicrystals and approximants acs by soft and hard x-ray photoemission spectroscopy

together the results offer a practical decision rule compute network metrics and interactions when local case

histories

case histories

together the results offer a practical decision rule compute network metrics and interactions when local case histories are coarse or delayed rely primarily on ar baselines when granular cases are timely using network signals mainly for diagnostic targeting

the street has emerged as a primary site where everyday publics are confronted with ai as an infrastructural phenomenon as machine learning-based

systems

ai assistance

the street has emerged as a primary site where everyday publics are confronted with ai as an infrastructural phenomenon as machine learning-based systems are now commonly deployed in this setting in the form of automated cars facial recognition smart billboards and the like

to overcome this barrier electricity from the main grid must be cleaner and cheaper and

energy

energy consumption

to overcome this barrier electricity from the main grid must be cleaner and cheaper and energy storage costs per unit stored must decrease

normative reasoning in large language models a comparative benchmark from

logical

normative reasoning

normative reasoning in large language models a comparative benchmark from logical and modal perspectives

inside core-kg evaluating structured prompting and

coreference

coreference resolution

inside core-kg evaluating structured prompting and coreference resolution for knowledge graphs