Blog
Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as Test-Time Scaling Evaluators
Scaling test-time computation, or affording a generator large language model (LLM) extra compute during inference, typically employs the help of external non-generative evaluators (i.e., reward models). Concurrently, LLM-judges, models trained to generate evaluations and critiques (explanations) in natural language, are becoming increasingly...
Impulsive pattern recognition of a myoelectric hand via Dynamic Time Warping
Although myoelectric prosthetic hands provide amputees with intuitive control, their reliance on many EMG sensors limits accessibility and makes them complex and expensive. To address this problem, this work presents a different perspective that makes use of a single EMG sensor and...
Broadband Fourier transform spectroscopy of quantum emitters photoluminescence with sub-nanosecond temporal resolution
The spectral characterization of quantum emitter luminescence over broad wavelength ranges and fast timescales is important for applications ranging from biophysics to quantum technologies. Here we present the application of time-domain Fourier transform spectroscopy, based on a compact and stable birefringent interferometer...
Joint Knowledge and Power Management for Secure Semantic Communication Networks
Recently, semantic communication (SemCom) has shown its great superiorities in resource savings and information exchanges. However, while its unique background knowledge guarantees accurate semantic reasoning and recovery, semantic information security-related concerns are introduced at the same time. Since the potential eavesdroppers may...
Diffusion Bridge Models for 3D Medical Image Translation
Diffusion tensor imaging (DTI) provides crucial insights into the microstructure of the human brain, but it can be time-consuming to acquire compared to more readily available T1-weighted (T1w) magnetic resonance imaging (MRI). To address this challenge, we propose a diffusion bridge model...
Addendum: Fitting the DESI BAO Data with Dark Energy Driven by the Cohen-Kaplan-Nelson Bound
Motivated by the recent Year-2 data release of the DESI collaboration, we update our results on time-varying dark energy models driven by the Cohen-Kaplan-Nelson bound. The previously found preference of time-dependent dark energy models compared to $\Lambda$CDM is further strengthend by the...
An LMM for Efficient Video Understanding via Reinforced Compression of Video Cubes
Large Multimodal Models (LMMs) uniformly perceive video frames, creating computational inefficiency for videos with inherently varying temporal information density. This paper present \textbf{Quicksviewer}, an LMM with new perceiving paradigm that partitions a video of nonuniform density into varying cubes using Gumbel Softmax,...
Efficient Testing Using Surrogate Information
In modern clinical trials, there is immense pressure to use surrogate markers in place of an expensive or long-term primary outcome to make more timely decisions about treatment effectiveness. However, using a surrogate marker to test for a treatment effect can be...
Stop Summation: Min-Form Credit Assignment Is All Process Reward Model Needs for Reasoning
Process reward models (PRMs) have proven effective for test-time scaling of Large Language Models (LLMs) on challenging reasoning tasks. However, reward hacking issues with PRMs limit their successful application in reinforcement fine-tuning. In this paper, we identify the main cause of PRM-induced...
A Short History of (Orbital) Decay: Roman’s Prospects for Detecting Dying Planets
The Roman Space Telescope Galactic Bulge Time Domain Survey (GBTDS) is expected to detect ~10^5 transiting planets. Many of these planets will have short orbital periods and are thus susceptible to tidal decay. We use a catalog of simulated transiting planet detections...