prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
respec towards optimizing speculative decoding in
reinforcement
continual learning
respec towards optimizing speculative decoding in reinforcement learning systems
this novel interdisciplinary approach addresses this gap by combining agent development and linguistic analysis to assess the prompted personality of llm-based
agents
ai literacy
this novel interdisciplinary approach addresses this gap by combining agent development and linguistic analysis to assess the prompted personality of llm-based agents in a poetry explanation task
acknowledging recent breakthroughs in the context of deep networks several architectural options have been deployed to analyze
supply
supply chain
acknowledging recent breakthroughs in the context of deep networks several architectural options have been deployed to analyze supply chain datasets
we investigate the semi-discrete optimal transport ot problem where a continuous source measure mu is transported to a discrete target
measure
optimal transport
we investigate the semi-discrete optimal transport ot problem where a continuous source measure mu is transported to a discrete target measure nu with particular attention to the ot map approximation
the source code used to conduct the experiments
presented
extensive experiments
the source code used to conduct the experiments presented in this paper is made freely available
multimodal large language models mllms exhibit a pronounced preference for textual inputs when processing vision-language
data
large language
multimodal large language models mllms exhibit a pronounced preference for textual inputs when processing vision-language data limiting their ability to reason effectively from visual evidence
our general interaction framework which reduces to several previously studied models as special cases provides a versatile platform for engineering quantum
correlations
quantum channels
our general interaction framework which reduces to several previously studied models as special cases provides a versatile platform for engineering quantum correlations with applications in quantum simulation state preparation and sensing protocols
to explore this possibility we reanalyse the region using observational molecular
line
absorption line
to explore this possibility we reanalyse the region using observational molecular line data
in particular we assume that the velocity field is induced by a finite set of cellular
flows
gradient flow
in particular we assume that the velocity field is induced by a finite set of cellular flows that can be controlled in time
we demonstrate both the generation and detection of an ultra-high flux of polarization bell
states
quantum technologies
we demonstrate both the generation and detection of an ultra-high flux of polarization bell states using broadband hyper-entangled bi-photons that are quantum-correlated in both polarization and time-energy
road traffic accidents remain a major public health challenge worldwide with
urbanisation
crash risk
road traffic accidents remain a major public health challenge worldwide with urbanisation and population density identified as key factors influencing risk
these results suggest that the density of the external medium regulates not only
protostellar
circumgalactic medium
these results suggest that the density of the external medium regulates not only protostellar mass growth but also the inflow-outflow balance and angular momentum transport in magnetized rotating star-forming cores
to reduce such bias debiased machine learning employs
neyman
debiased machine learning
to reduce such bias debiased machine learning employs neyman orthogonal estimating equations
focusing on a reduced and hardware-feasible version of the mnist digit classification task or near-term photonic processors it offers a concrete framework to evaluate how
photonic
photonic crystal
focusing on a reduced and hardware-feasible version of the mnist digit classification task or near-term photonic processors it offers a concrete framework to evaluate how photonic quantum circuits learn and generalize from limited data
steervlm robust model control through lightweight activation steering for vision
language
vision transformers
steervlm robust model control through lightweight activation steering for vision language models
using the derived mathematical tools we assess the impact of these two control actions and determine the optimal
control
optimal control
using the derived mathematical tools we assess the impact of these two control actions and determine the optimal control policy
crucially we uncover an emergent capability for instruction-based decoding control the model learns to interpret
natural
language models
crucially we uncover an emergent capability for instruction-based decoding control the model learns to interpret natural language commands e
these findings suggest that resting-state
eeg
brain-computer interface
these findings suggest that resting-state eeg connectivity patterns can index stable cognitive traits such as creativity
to address this challenge we propose a novel strategy called inference-aware policy optimization which modifies policy optimization to account for how the
policy
policy optimization
to address this challenge we propose a novel strategy called inference-aware policy optimization which modifies policy optimization to account for how the policy will be evaluated downstream
for fluorescence-based readout we selectively couple one of the t center spin-conserving transitions to a single-mode
photonic
single photons
for fluorescence-based readout we selectively couple one of the t center spin-conserving transitions to a single-mode photonic cavity exploiting the enhancement of the fluorescence emission and cyclicity
082 thereby confirming a superior balance between
image
depth estimation
082 thereby confirming a superior balance between image fidelity and depth accuracy
using these benchmarks researchers have found that
llms
llm reasoning
using these benchmarks researchers have found that llms sometimes fail to keep secrets private when responding to complex tasks e
the advent of artificial intelligence ai -native wireless communication is fundamentally reshaping the design paradigm of next-generation nextg systems where
intelligent
reconfigurable intelligent
the advent of artificial intelligence ai -native wireless communication is fundamentally reshaping the design paradigm of next-generation nextg systems where intelligent air interfaces are expected to operate adaptively and efficiently in highly dynamic environments
single mode lasing and spectral narrowing in
photonic
waveguide modes
single mode lasing and spectral narrowing in photonic crystal line-defect cavities via spatially selected bloch modes
we release amo-bench to facilitate further research into advancing the reasoning
abilities
language agents
we release amo-bench to facilitate further research into advancing the reasoning abilities of language models
self-supervised learning ssl holds a great deal of promise for
applications
artificial neural
self-supervised learning ssl holds a great deal of promise for applications in neuroscience due to the lack of large-scale consistently labeled neural datasets
interstellar comet 3i atlas evidence for galactic
cosmic
quiescent galaxies
interstellar comet 3i atlas evidence for galactic cosmic ray processing
we propose that atomic hydrogen facilitates the formation of radical sites which promotes covalent bond formation between adjacent particles or
molecular
molecular dynamics
we propose that atomic hydrogen facilitates the formation of radical sites which promotes covalent bond formation between adjacent particles or molecular units creating a more interconnected and rigid network with smaller interlayer distance
this progressive learning journey guides agents from mastering fundamental flight control to executing sophisticated cooperative
multi-drone
control strategy
this progressive learning journey guides agents from mastering fundamental flight control to executing sophisticated cooperative multi-drone racing strategies
we further show that abstention is a versatile tool with direct applications to other core problems in policy learning it yields improved guarantees under margin conditions without the common realizability assumption connects to distributionally robust policy learning by hedging against small data shifts and supports s...
policy
policy optimization
we further show that abstention is a versatile tool with direct applications to other core problems in policy learning it yields improved guarantees under margin conditions without the common realizability assumption connects to distributionally robust policy learning by hedging against small data shifts and supports s...
to mitigate the downlink peak-to-average power ratio increase caused by fdma we further develop a multi-user downlink ce
transmission
transmit power
to mitigate the downlink peak-to-average power ratio increase caused by fdma we further develop a multi-user downlink ce transmission scheme including multiple access mechanism downlink control information design and corresponding system-level implementation which ensures compatibility with the new radio standard
experimental verification is scarce however especially in the telecom range because real
photonic
photonic devices
experimental verification is scarce however especially in the telecom range because real photonic crystals and experimental methods inherently cannot be homogeneous in the third dimension
these findings suggest that resting-state eeg connectivity patterns can index stable
cognitive
human cognition
these findings suggest that resting-state eeg connectivity patterns can index stable cognitive traits such as creativity
enhancing the reachability of variational quantum
algorithms
quantum advantage
enhancing the reachability of variational quantum algorithms via input-state design
we introduce debate2create d2c a framework in which large language model llm
agents
learning agents
we introduce debate2create d2c a framework in which large language model llm agents engage in a structured dialectical debate to jointly optimize a robot s design and its reward function
we propose link-kg a modular framework that integrates a three-stage llm-guided
coreference
coreference resolution
we propose link-kg a modular framework that integrates a three-stage llm-guided coreference resolution pipeline with downstream kg extraction
beating the winner s curse via inference-aware
policy
policy optimization
beating the winner s curse via inference-aware policy optimization
bayesian nonlinear pde inference via gaussian
process
gaussian process
bayesian nonlinear pde inference via gaussian process collocation with application to the richards equation
photo-induced force microscopy shows that introducing tin into the van der waals gap modulates the
phonon
phonon polaritons
photo-induced force microscopy shows that introducing tin into the van der waals gap modulates the phonon polariton dispersion by up to 38
along the way we extend previous error analyses of brenier maps to the case of conditional brenier maps that arise in the context of
simulation
simulation studies
along the way we extend previous error analyses of brenier maps to the case of conditional brenier maps that arise in the context of simulation based inference
several researchers introduced different methods to decompose n 1 -bit toffoli gates in a quantum circuit into a set of standard 3-bit toffoli gates or a set of elementary quantum
gates
-bit toffoli gates
several researchers introduced different methods to decompose n 1 -bit toffoli gates in a quantum circuit into a set of standard 3-bit toffoli gates or a set of elementary quantum gates such as single-qubit and two-qubit gates
theoretical methods for computing the band structure in 2d are well-established and fast because 2d
photonic
photonic circuits
theoretical methods for computing the band structure in 2d are well-established and fast because 2d photonic crystals are homogeneous in the third dimension
modeling object attention in mobile ar for intrinsic
cognitive
mobile ar
modeling object attention in mobile ar for intrinsic cognitive security
this work provides a unified framework for examining how individual learnability and structural adaptability drive the formation of socioeconomic networks with diverse topologies offering
new
complex networks
this work provides a unified framework for examining how individual learnability and structural adaptability drive the formation of socioeconomic networks with diverse topologies offering new insights into adaptive behavior systemic organization and resilience
one of the most fundamental questions in laser physics is the following which
mode
waveguide modes
one of the most fundamental questions in laser physics is the following which mode of an optical cavity will reach the lasing threshold first when gain is applied
in this paper we systematically evaluate llms
reasoning
large language models llms
in this paper we systematically evaluate llms reasoning capabilities in the normative domain from both logical and modal perspectives
model and hyperparameter selection are critical but challenging in
machine
machine learning
model and hyperparameter selection are critical but challenging in machine learning typically requiring expert intuition or expensive automated search
the system features a counter-propagating beam at the same wavelength as the
quantum
quantum technologies
the system features a counter-propagating beam at the same wavelength as the quantum state which simultaneously actively stabilizes the cavity and after transmission acts as the local oscillator for homodyne detection
our mv-mlm leverages multi-view supervision to learn rich
representations
representation learning
our mv-mlm leverages multi-view supervision to learn rich representations from extensive radiology data by employing cross-modal self-supervision across image-text pairs
we apply inputdsa on recurrent neural networks rnns
trained
neural networks
we apply inputdsa on recurrent neural networks rnns trained with deep reinforcement learning identifying that high-performing networks are dynamically similar to one another while low-performing networks are more diverse
the resulting half-light radii are in good agreement with literature values but
reveal
interstellar medium
the resulting half-light radii are in good agreement with literature values but reveal subtle variations linked to differences in stellar distribution and morphology
our analysis allows for estimating a diverging number of treatment effects simultaneously and establishes the consistency and asymptotic normality of the
regression-based
average treatment effect
our analysis allows for estimating a diverging number of treatment effects simultaneously and establishes the consistency and asymptotic normality of the regression-based estimators
consequently prema poses a novel threat to the integrity of multi-modal
diffusion
diffusion models
consequently prema poses a novel threat to the integrity of multi-modal diffusion models particularly in image-editing applications that operate with fixed prompts
the observed polarizabilities are 3-25 times larger than those of tin vacancy centers which we attribute to valence
band
photonic crystal
the observed polarizabilities are 3-25 times larger than those of tin vacancy centers which we attribute to valence band resonances that delocalize the e_u wavefunctions
we investigate a quantum heat engine where energy exchanges are driven by generalized measurements and the sequence of these
operations
quantum channels
we investigate a quantum heat engine where energy exchanges are driven by generalized measurements and the sequence of these operations is coherently controlled in a superposition of causal orders
our findings reveal how families co-construct children s ai literacy exposing tensions between practical expectations and
critical
ai use
our findings reveal how families co-construct children s ai literacy exposing tensions between practical expectations and critical literacies and provide design implications that foster sdl while balancing autonomy and oversight
we find that dwarf agn selected by infrared colors are the most distinct
population
stellar mass
we find that dwarf agn selected by infrared colors are the most distinct population with the highest star formation rates and lowest stellar masses
results show lower rmse and superior cyclical responsiveness for fpas zeta along with early-warning capability for shocks and regime shifts indicating practical value for
policy
monetary policy
results show lower rmse and superior cyclical responsiveness for fpas zeta along with early-warning capability for shocks and regime shifts indicating practical value for policy institutions
since the early 1990s the field has evolved through distinct historical phases each shaped by the intelligence level of machines from early human--computer interaction frameworks built around primitive computers to today s human--agent interaction paradigms driven by intelligent agents and potentially to human--level o...
intelligence
artificial intelligence
since the early 1990s the field has evolved through distinct historical phases each shaped by the intelligence level of machines from early human--computer interaction frameworks built around primitive computers to today s human--agent interaction paradigms driven by intelligent agents and potentially to human--level o...
2024 develops riesz regression for automatic debiased machine learning which directly estimates the
riesz
riesz regression
2024 develops riesz regression for automatic debiased machine learning which directly estimates the riesz representer or equivalently the bias-correction term by minimizing the mean squared error
reinforcement learning rl is widely used to produce robust robotic
manipulation
imitation learning
reinforcement learning rl is widely used to produce robust robotic manipulation policies but fine-tuning vision-language-action vla models with rl can be unstable due to inaccurate value estimates and sparse supervision at intermediate steps
we establish the convergence of the sequence generated by tkma to a fixed
point
superlinear convergence
we establish the convergence of the sequence generated by tkma to a fixed point of t
in contrast adjacent generative fields most notably video
generation
video generation
in contrast adjacent generative fields most notably video generation vigen have demonstrated remarkable generalization in modeling human behaviors highlighting transferable insights that mogen can leverage
while imitation learning il enables effective visual
navigation
visual navigation
while imitation learning il enables effective visual navigation il policies are prone to unpredictable failures in out-of-distribution ood scenarios
we also provide new data-driven control design methods in terms of
linear
data-driven stabilization
we also provide new data-driven control design methods in terms of linear matrix inequalities that complement the conditions for informativity
in the experimental section we present a unique dataset containing high-trial inner speech
eeg
electroencephalography eeg
in the experimental section we present a unique dataset containing high-trial inner speech eeg meg and preliminary optically pumped magnetometer opm data
a structured workflow was used to clean and prepare
data
data collection
a structured workflow was used to clean and prepare data for analysis
the evaluation examines multiple llms under varied input specifications
retrieval-augmented
models llms
the evaluation examines multiple llms under varied input specifications retrieval-augmented configurations and documentation completeness levels
our work establishes a connection between proper scoring rules and menu design showing how the structure of the hypothesis
test
test statistics
our work establishes a connection between proper scoring rules and menu design showing how the structure of the hypothesis test constrains the elicitable information
we present the first large scale benchmark of 48 deep learning optical flow models on radarsat 2 scansar sea
ice
sea ice
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
little is known about how this ability is affected by simultaneously present
distractor
visual stimuli
little is known about how this ability is affected by simultaneously present distractor sounds despite the real-life importance of tracking moving sounds in the presence of distractors
we investigate a quantum heat engine where
energy
quantum walk
we investigate a quantum heat engine where energy exchanges are driven by generalized measurements and the sequence of these operations is coherently controlled in a superposition of causal orders
in contrast to the existing approaches we regard classification
training
federated learning
in contrast to the existing approaches we regard classification training as a sequential process where classes are learned sequentially which we call emph inductive approach
the system can determine whether real-time behavior is
consistent
real-world applications
the system can determine whether real-time behavior is consistent with expected commands
we highlight critical open questions and practical trade-offs that must be addressed offering multidisciplinary insights from engineering logic and law to guide future developments in legally compliant autonomous
driving
autonomous driving
we highlight critical open questions and practical trade-offs that must be addressed offering multidisciplinary insights from engineering logic and law to guide future developments in legally compliant autonomous driving systems
our findings pave the way for the integration of chemically tailored intercalation compounds in scalable
quantum
quantum technologies
our findings pave the way for the integration of chemically tailored intercalation compounds in scalable quantum technologies
43 demonstrating superior capability in detecting anomalies accurately and promptly as compared to existing classical and
quantum
anomaly detection
43 demonstrating superior capability in detecting anomalies accurately and promptly as compared to existing classical and quantum models
this amplified b-field is likely sustained by the turbulent kinetic energy in the outflow and may be critical in directing the transport of metals and cosmic rays into the
circumgalactic
circumgalactic medium
this amplified b-field is likely sustained by the turbulent kinetic energy in the outflow and may be critical in directing the transport of metals and cosmic rays into the circumgalactic medium
likelihood-free inference of phylogenetic
tree
phylogenetic tree
likelihood-free inference of phylogenetic tree posterior distributions
we present a generalization of this problem where inspired by biological considerations the
food
food web
we present a generalization of this problem where inspired by biological considerations the food web has weighted edges to represent the importance of predator-prey relationships
this paper proposes neurodob a deep neural network based observer controller for vehicle lateral dynamics which replaces the conventional disturbance
observer
disturbance observer
this paper proposes neurodob a deep neural network based observer controller for vehicle lateral dynamics which replaces the conventional disturbance observer dob with a deep neural network dnn to enhance personalized lateral control
stakeholders across 21 european islands were consulted on
climate
climate change
stakeholders across 21 european islands were consulted on climate and land use change issues affecting ecosystem services
taking a solution-centric approach we introduce the concept of provably small set of solutions p called a it portfolio such that for every objective function h cdot in the given class mathbf c there exists some solution in
p
-approximation algorithm
taking a solution-centric approach we introduce the concept of provably small set of solutions p called a it portfolio such that for every objective function h cdot in the given class mathbf c there exists some solution in p which is an alpha -approximation for h cdot
bayesian network fusion of large language models for
sentiment
large language models llms
bayesian network fusion of large language models for sentiment analysis
given a stream of asynchronous and spatially sparse events our goal is to discover an
event
temporal understanding
given a stream of asynchronous and spatially sparse events our goal is to discover an event representation that preserves these properties
here we show that lyman- alpha ly alpha radiation pressure limits the gas mass converted into
stars
star formation
here we show that lyman- alpha ly alpha radiation pressure limits the gas mass converted into stars particularly in primordial environments
turbulence induced photon statistics with classical beam propagation in free
space
optical communication
turbulence induced photon statistics with classical beam propagation in free space optical communications
finally drawing on literature from both nlp ml and hci as a constructive step forward we develop reflection prompts to support hci practitioners engage with
llm
llm reasoning
finally drawing on literature from both nlp ml and hci as a constructive step forward we develop reflection prompts to support hci practitioners engage with llm reasoning in an informed and reflective way
analyzing parametric oscillator ising machines through the
kuramoto
parametric oscillator
analyzing parametric oscillator ising machines through the kuramoto lens
hybrid dqn-td3 reinforcement learning for
autonomous
reinforcement learning
hybrid dqn-td3 reinforcement learning for autonomous navigation in dynamic environments
see4d pose-free 4d generation via auto-regressive
video
optical flow
see4d pose-free 4d generation via auto-regressive video inpainting
insect species subject to infection predation and
anisotropic
butterfly species
insect species subject to infection predation and anisotropic environmental conditions may exhibit preferential movement patterns
allowing the order of quantum operations to exist in superposition is known to
open
quantum computing
allowing the order of quantum operations to exist in superposition is known to open new routes for thermodynamic tasks
we show that the system is approximately controllable for blowup points with feedback controls and the feedback operator is
bounded
linear control
we show that the system is approximately controllable for blowup points with feedback controls and the feedback operator is bounded at any time before blowup
to overcome this limitation we propose to generalize sobolev ipm through the lens of emph orlicz geometric structure which employs convex functions to capture nuanced
geometric
gradient flow
to overcome this limitation we propose to generalize sobolev ipm through the lens of emph orlicz geometric structure which employs convex functions to capture nuanced geometric relationships building upon recent advances in optimal transport theory -- particularly orlicz-wasserstein ow and generalized sobolev transport...
in this setting our second algorithm a 2 -off-c 2 pl actively selects samples that target the least-informative
dimensions
preference learning
in this setting our second algorithm a 2 -off-c 2 pl actively selects samples that target the least-informative dimensions of the test user s preference
large language models llms often struggle with complex
mathematical
mathematical reasoning
large language models llms often struggle with complex mathematical reasoning where prose-based generation leads to unverified and arithmetically unsound solutions
our experiments show that current video large language model video-llm architectures have critical limitations in temporal understanding struggling with tasks that require detailed comprehension of action sequences and
temporal
temporal understanding
our experiments show that current video large language model video-llm architectures have critical limitations in temporal understanding struggling with tasks that require detailed comprehension of action sequences and temporal progression
estimating unrestricted spatial interdependence in
panel
panel data
estimating unrestricted spatial interdependence in panel spatial autoregressive models with latent common factors
biaxial anisotropy arising from distinct optical responses along three principal directions underlies the
complex
optical properties
biaxial anisotropy arising from distinct optical responses along three principal directions underlies the complex structure of many crystalline polymeric and biological materials