gender
PulseAugur coverage of gender — every cluster mentioning gender across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
AI systems continue to struggle with race and gender bias
AI systems continue to exhibit significant biases related to race and gender, a persistent issue that has been documented and discussed. These biases manifest in various AI applications, raising concerns about fairness …
-
Learner errors amplify social bias in large language models, study finds
A new study from Hugging Face explores how grammatical errors in prompts affect social bias in large language models. Researchers found that introducing common learner errors, particularly with prepositions, significant…
-
Activation steering improves ASR performance and fairness
Researchers have explored activation steering as a method to enhance both performance and fairness in Automatic Speech Recognition (ASR) systems. Their investigation focused on two techniques: paired activation subtract…
-
Ex-Palantir execs raise $17M for AI construction design startup Dream
A new startup called Dream, founded by former Palantir executives, has secured $17 million in seed funding led by Andreessen Horowitz. The company aims to revolutionize the architecture, engineering, and construction (A…
-
New challenge targets face-voice models' reliance on language and gender
A new challenge, the Face-voice Association across LAnguages and Gender (FLAG) 2027 Challenge, has been introduced to address limitations in current face-voice association models. These models often rely on language or …
-
US designates Italian activist collective Autistici/Inventati as terrorist org
The Italian collective Autistici/Inventati, which provided digital services for global NGOs and activist groups for 25 years, was designated a terrorist organization by the U.S. Department of State and Treasury on Augus…
-
Internet and Society Center study days move online due to high demand
The Internet and Society Center is holding study days, with attendance exceeding in-person capacity, necessitating a shift to virtual attendance. These events cover a range of topics including AI, generative AI, gender,…
-
AI integration highlights broken workforce supply chains, demanding better talent data
The integration of AI into the workforce presents a significant challenge due to a broken supply chain for talent data. Businesses need to treat their workforce like a supply chain, similar to how Walmart managed its re…
-
Lumière Summit adopts 5 declarations on AI, piracy, gender, education, environment · 3 sources tracked
The Lumière Summit has concluded with the adoption of five declarations concerning artificial intelligence, piracy, gender, education, and the environment within the screen industries. These declarations were agreed upo…
-
Multimodal LLMs Predict Urban Safety but Show Demographic Bias
A new research paper explores the capabilities of Multimodal Large Language Models (MLLMs) in assessing perceived urban safety from street-view imagery. While these models demonstrate a zero-shot capability to predict s…
-
LLMs show bias toward majority groups when prompted with demographics
A new study has revealed that large language models (LLMs) do not act as neutral judges when prompted with demographic information. Instead, models without any demographic conditioning tend to align with the judgments o…
-
Survey details periocular biometrics for media forensics and disinformation detection
A new survey paper explores the use of periocular soft biometrics for applications in multimedia forensics and disinformation detection. The paper details demographic attribute estimation from periocular images, coverin…
-
AI continues to struggle with race and gender bias
AI systems continue to exhibit significant biases related to race and gender, a problem that persists despite ongoing development. These biases manifest in various AI applications, highlighting a critical challenge for …
-
New method improves concept unlearning in LLMs
Researchers have developed MPSelectTune, a novel method to improve concept unlearning in large language models (LLMs). This approach uses a multi-task loss and adversarial fine-tuning, focusing on the prompt type that y…
-
New benchmark evaluates LALMs on nuanced speech characteristics · 2 sources tracked
Researchers have introduced ParaPairAudioBench, a new benchmark designed to evaluate Large Audio-Language Models (LALMs) in their ability to distinguish fine-grained paralinguistic features in speech. The benchmark comp…
-
New ML framework tackles intersectional bias with coverage constraints
Researchers have developed a new framework to mitigate bias in machine learning models, particularly for individuals at the intersection of multiple sensitive attributes like race and gender. This approach incorporates …
-
Christian cell network launches with default blocks on porn and LGBTQ+ content
A new US cell phone network called Radiant Mobile, targeting Christians, is set to launch next month with a focus on blocking explicit content. The network will implement network-level blocking for pornography and other…