Claims Browser

Conventional frame-based computer vision approaches have practical flaws including size, weight, power consumption constraints, security concerns with cloud computation, transmission latency, and connectivity requirements

"This approach has several practical flaws. The size, weight and power consumption of the device could prohibit deployment at the extreme edge or in covert sensing environments. Besides this, there are security concerns inherent in cloud-based or other off-device computation approaches due to the requirement of sending and receiving potentially sensitive data. Furthermore, this transmission of data introduces latency and requires consistent connectivity to the cloud infrastructure to function."

Vision Language Models still struggle to deeply understand text within images despite their success in general visual tasks

"Vision Language Models (VLMs) have shown great success in general visual tasks, yet they still struggle to deeply understand text within images."

There is a striking performance gap between even the best VLMs and humans on visual text error correction tasks

"Our experiments reveal a striking performance gap between even the best VLMs and human"

Most VLMs struggle to accurately perceive visual text, resulting in frequent correction errors

"most models struggle to accurately perceive the visual text, resulting in frequent correction errors"

Visual text error correction requires profound understanding of the interplay between visual text and its surrounding visual context

"which requires a profound understanding of the interplay between visual text and its surrounding visual context"

Existing reinforcement-learning methods for robot crowd navigation are limited because they output a single reactive action at each timestep, which constrains their ability to represent diverse short-term avoidance strategies

"Existing reinforcement-learning methods typically output a single reactive action at each timestep, which limits their ability to represent diverse short-term avoidance strategies."

Action chunks are especially important for crowd navigation in bounded environments where boundary violations are treated as collisions

"ablations demonstrate that action chunks are especially important for the modified bounded benchmark"

Common crowd-navigation benchmarks have an evaluation artifact where learned agents can leave the valid domain and bypass dense crowds without explicit boundary constraints

"we observe an evaluation artifact in common crowd-navigation benchmarks: without explicit boundary constraints, learned agents may leave the valid domain and bypass dense crowds"

Document-to-document LLM translation in a single pass frequently suffers from structural misalignment including sentence omissions or hallucinations

"document-to-document generation in a single pass frequently suffers from structural misalignment, manifesting as sentence omissions or hallucinations that violate the core requirement of source-target correspondence"

StarPO framework significantly enhances translation quality and structural integrity in document-level machine translation

"Experimental results across news and literary domains demonstrate that StarPO significantly enhances translation quality and structural integrity"

Compact models using StarPO can surpass GPT-4o performance while maintaining superior token efficiency

"StarPO allows compact models to surpass the performance of massive proprietary systems like GPT-4o while maintaining superior token efficiency"

Autoregressive large language models routinely generate factually incorrect outputs with high decoding confidence, limiting their deployment in high-stakes workflows

"Autoregressive large language models (LLMs) routinely generate factually incorrect outputs with high decoding confidence, limiting their deployment in high-stakes workflows."

Existing output-stage uncertainty metrics can fail when models are overconfident on false assertions

"Existing output-stage uncertainty metrics can fail when models are overconfident on false assertions"

Internal hidden-state transition dynamics during generation can signal factual errors without auxiliary decoding calls

"internal hidden-state transition dynamics during generation can signal factual errors without auxiliary decoding calls"

PoP operates within the base forward pass, adding less than 1.2% runtime latency and requiring zero additional generation passes

"The mechanism operates within the base forward pass, adding less than 1.2% runtime latency and requiring zero additional generation passes."

LLM agents commit to directional calls on unpredictable questions 54.0% of the time when shown professional-looking market panels, compared to only 6.5% when asked bare questions, across 12 frontier models

"An LLM agent shown a professional-looking market panel commits to a directional call on a provably unpredictable question far more often than one asked the bare question: across 12 frontier models, commitment rises from 6.5% to 54.0% as evidence is escalated."

LLMs commit to unpredictable questions just as readily when all data is fabricated, with commitment rising from 24.5% to 36.8% with fake data versus 37.6% with genuine data

"It commits just as readily when every number on the panel is invented: fabricating the entire display, so nothing the model can see is true except the question itself, still lifts commitment from 24.5% to 36.8%, statistically indistinguishable from the 37.6% produced by genuine market data."

What drives confident LLM action is not information content but the authority of information packaging

"What unlocks confident action is not information but the authority of its packaging."