Bowen Jin, Gang Liu, Chi Han, Meng Jiang, Heng Ji and Jiawei Han explore the potential of large language models (LLMs) in processing text data associated with rich structure information or graph data with textual information, and their reasoning ability in graph scenarios.

Albert Gu and Tri Dao introduce Mamba, a neural network architecture that uses selective state space models (SSMs) to model long sequences. This approach addresses the weaknesses of other subquadratic-time architectures in handling content-based reasoning for discrete modalities such as language. Mamba outperforms Transformer...

Meta introduce Seamless Communication, a new system for expressive cross-lingual communication in real time. It includes SeamlessExpressive, which preserves expression in speech-to-speech translation, and SeamlessStreaming, a streaming translation model with low latency.

Tobias Kirschstein, Simon Giebenhain, and Matthias Nießner propose DiffusionAvatars, a diffusion-based neural renderer that generates high-fidelity 3D head avatars with intuitive control over pose and expression. Trained on RGB videos and corresponding tracked NPHM meshes of a person, DiffusionAvatars outperforms existing app...

DeepMind have discovered 2.2 million new crystals using their AI tool GNoME, with 380,000 being stable and potentially viable for future technologies. This discovery has the potential to drive research in various fields including superconductors, powering supercomputers, next-generation batteries and electric vehicles.

Pavan Kumar Anasosalu Vasu, Hadi Pouransari, Fartash Faghri, Raviteja Vemulapalli, and Oncel Tuzel introduce MobileCLIP, a new family of efficient image-text models optimized for runtime performance. The proposed training approach, multi-modal reinforced training, leverages knowledge transfer from an image captioning model an...

Han Altae-Tran, Soumya Kannan , Anthony J Suberski, Kepler S Mears, F Esra Demircioglu, Lukas Moeller, Selin Kocalar, Rachel Oshiro , Kira S Makarova , Rhiannon K Macrae , Eugene V Koonin, Feng Zhang identified 188 previously unreported CRISPR-linked gene modules, revealing many additional biochemical functions coupled to ada...

Financial Times - Ardian, a leading European private equity firm, faces a succession problem. The firm's founder, Dominique Senequier, has yet to announce her retirement plans.

Authors introduce Orca 2, a model that enhances smaller language models' reasoning abilities by using various training signals. It outperforms similar-sized models and achieves performance close to larger ones on complex tasks.

Quantamagazine - Explore the fascinating world of recursive sequences with mathematicians Alex Stone and Kristina Armitage. Learn about Michael Somos' counterintuitive sequence and Fritz Göbel's surprising integer trends. The puzzling patterns have revealed ties to graph theory and prime numbers, awing mathematicians.

Pew Research Center investigate how Americans consume news on social media, revealing that Facebook leads the way with 30% of U.S. adults regularly getting news there. Demographic differences and partisan divides emerge in platform preferences.

Wired - The story of three young hackers, Josiah White, Paras Jha and Dalton Norman, who built a powerful web-attack tool called Mirai. The tool was used in some of the most disruptive cyberattacks of recent years.

VentureBeat - Two weeks before OpenAI released ChatGPT — Meta released a research demo called Galactica but had to take it down after an outcry over what was, back then, a word that had not yet made it into the mainstream: Hallucinations.

VIDEO | Jian Tan (Alibaba) presents a domain knowledge augmented AI for Databases at Carnegy Mellon Unversity

IPVM report on a PRC 'Smart Campus' system that alerts when ethnic minority students are suspected of fasting for Ramadan

Charles Frye discusses the intersection of large language models (LLMs) and operating system kernels. He highlights recent research that applies systems metaphors to improve LLMs, focusing on pretraining methods, inference-time speed optimizations, and prompting strategies

Jiazhan Feng, Ruochen Xu, Junheng Hao, Hiteshi Sharma, Yelong Shen, Dongyan Zhao and Weizhu Chen introduce LoGiPT, a novel language model that directly emulates the reasoning processes of logical solvers. LoGiPT outperforms state-of-the-art solver-augmented LMs and few-shot prompting methods on competitive LLMs like ChatGPT o...

Maxine (aka crumb) presents “Reweighting,” a novel fine-tuning technique, epitomizes this dual focus, offering a method to enhance model performance through very short continued pretraining with a minimal dataset.

Shilong Liu, Hao Cheng, Haotian Liu, Hao Zhang, Feng Li, Tianhe Ren, Xueyan Zou, Jianwei Yang, Hang Su, Jun Zhu, Lei Zhang, Jianfeng Gao, and Chunyuan Li present LLaVA-Plus, a multimodal assistant that extends LLaVA by incorporating external tools for real-world tasks. The system achieves new SoTA on VisIT-Bench with consiste...

Shuo Yang, Wei-Lin Chiang, Lianmin Zheng, Joseph E. Gonzalez and Ion Stoica examine the trustworthiness of public benchmarks for large language models due to potential contamination in pre-training or fine-tuning datasets. They demonstrate that simple variations of test data, such as paraphrasing and translation, can easily b...

Archiki Prasad, Alexander Koller, Mareike Hartmann, Peter Clark, Ashish Sabharwal, Mohit Bansal and Tushar Khot introduce ADaPT, a new approach for using large language models (LLMs) in interactive decision-making tasks requiring planning and adaptation. ADaPT outperforms established baselines by explicitly planning and recur...

Qinghao Ye, Haiyang Xu, Jiabo Ye, Ming Yan, Anwen Hu, Haowei Liu, Qi Qian, Ji Zhang, Fei Huang, Jingren Zhou introduce mPLUG-Owl2, a versatile multi-modal large language model that leverages modality collaboration to improve performance in both text and multi-modal tasks. The model utilizes a modularized network design and ac...

Nature - Viren Jain explains that early machine-learning systems were inspired by neural networks — now AI might allow neuroscientists to get to grips with the brain’s unique complexities.

Vitalik shares his warm, yet nuanced, feelings about techno-optimism

Alex Bauerle used Zeno to dive into the data and explore what the benchmark actually measures. What tasks does it test? What does the data look like?