国际著名电化学家、美国国家科学院院士、美国艺术与科学院院士、德克萨斯大学奥斯汀分校化学系教授Allen Joseph Bard博士于2024年2月11日与世长辞,享年90岁。Bard教授于1933年出生于美国纽约,获纽约城市学院学士学位、哈佛大学博士学位,1958年入职德克萨斯大学奥斯汀分校,担任化学系教授65年,被誉为校史上最重要的科学家。
Google's Bard has emerged as a formidable competitor to OpenAI's ChatGPT in the field of conversational AI.Notably,Bard has recently been updated to handle visual inputs alongside text prompts during conversations.Given Bard's impressive track record in handling textual inputs,we explore its capabilities in understanding and interpreting visual data(images)conditioned by text questions.This exploration holds the potential to unveil new insights and challenges for Bard and other forthcoming multi-modal Generative models,especially in addressing complex computer vision problems that demand accurate visual and language understanding.Specifically,in this study,we focus on 15 diverse task scenarios encompassing regular,camouflaged,medical,under-water and remote sensing data to comprehensively evaluate Bard's performance.Our primary finding indicates that Bard still struggles in these vision scenarios,highlighting the significant gap in vision-based understanding that needs to be bridged in future developments.We expect that this empirical study will prove valuable in advancing future models,leading to enhanced capabilities in comprehending and interpreting finegrained visual data.Our project is released on https://github.com/htqin/GoogleBard-VisUnderstand.
Haotong QinGe-Peng JiSalman KhanDeng-Ping FanFahad Shahbaz KhanLuc Van Gool
This study evaluated three prominent Large Language Models(LLMs)-Google’s AI BARD,Bing’s AI,and ChatGPT-4 in providing patient advice for hand laceration.Five simulated patient inquiries on hand trauma were prompted to them.A panel of Board-certified plastic surgical residents evaluated the responses for accuracy,comprehensiveness,and appropriate sources.Responses were also compared against existing literature and guidelines.This study suggests that ChatGPT outperforms BARD and Bing AI in providing reliable,evidence-based clinical advice,but they still face limitations in depth and specificity.Healthcare professionals are essential in interpreting LLM recommendations,and future research should improve LLM performance by integrating specialized databases and human expertise to advance nerve injury management and optimize patient-centred care.
Bryan LimIshith SethGabriella BullochYi XieDavid J Hunter-SmithWarren M Rozen
Radiotherapy is widely used in the management of advanced colorectal cancer(CRC).However,the clinical efficacy is limited by the safe irradiated dose.Sensitizing tumor cells to radiotherapy via interrupting DNA repair is a promising approach to conquering the limitation.The BRCA1-BARD1 complex has been demonstrated to play a critical role in homologous recombination(HR)DSB repair,and its functions may be affected by HERC2 or BAP1.Accumulated evidence illustrates that the ubiquitination-deubiquitination balance is involved in these processes;however,the precise mechanism for the cross-talk among these proteins in HR repair following radiation hasn’t been defined.Through activity-based profiling,we identified PT33 as an active entity for HR repair suppression.Subsequently,we revealed that BAP1 serves as a novel molecular target of PT33 via a CRISPR-based deubiquitinase screen.Mechanistically,pharmacological covalent inhibition of BAP1 with PT33 recruits HERC2 to compete with BARD1 for BRCA1 interaction,interrupting HR repair.Consequently,PT33 treatment can substantially enhance the sensitivity of CRC cells to radiotherapy in vitro and in vivo.Overall,these findings provide a mechanistic basis for PT33-induced HR suppression and may guide an effective strategy to improve therapeutic gain.
