Keno Jüchems
Lead Machine Learning Scientist at Limbic.
Oxford, England, United Kingdom
447 followers
414 connections
About
I am a Machine Learning Scientist at Limbic. I use techniques from deep learning, NLP, and machine learning to help improve patient outcomes in mental health.
Previously, I was a postdoc at the University of Oxford (Junior Research Fellow at St John’s College) and before that a DPhil student, where I investigated how people make complex decisions and set themselves goals.
Throughout my career I have used computational tools from cognitive science, psychology, economics, and machine learning, e.g. deep RL, functional magnetic resonance imaging, or microeconomic models such as prospect theory.
Activity
-
I am excited to be speaking at the GenAI Summit in Santa Clara this weekend (Nov 1-3). Very much looking forward to contribute to the expert panel…
I am excited to be speaking at the GenAI Summit in Santa Clara this weekend (Nov 1-3). Very much looking forward to contribute to the expert panel…
Liked by Keno Jüchems
-
We’re hiring! Come work with us in the Research Team at Limbic as an ML Scientist. If you’re looking to apply your ML skills and clinical domain…
We’re hiring! Come work with us in the Research Team at Limbic as an ML Scientist. If you’re looking to apply your ML skills and clinical domain…
Shared by Keno Jüchems
-
Such a pleasure speaking at this year’s BABCP conference alongside Graham Thew & Chloe Chessell in our symposium on implementation of digital…
Such a pleasure speaking at this year’s BABCP conference alongside Graham Thew & Chloe Chessell in our symposium on implementation of digital…
Liked by Keno Jüchems
Other similar profiles
Explore more posts
-
Federico Tartarini
In this video, I dive deep into using loops and conditional statements (if statements) to streamline your workflow and add dynamic content to your LaTeX documents. Whether you're a seasoned LaTeX user or just getting started, this video will equip you with the skills to: • Automatically generate numbered lists and figures • Include figures and captions based on conditions • Create dynamic graphics using loops and TikZ Check out the video here: https://lnkd.in/gc-u8TFp #latex #loops #conditionals #automation #productivity #scientificwriting #research #phdlife #engineering #mathematics
-
Merve Noyan
Chameleon 🦎 by Meta is now available in Hugging Face transformers 😍 A vision language model that comes in 7B and 34B sizes 🤩 But what makes this model so special? Demo and more in comments, keep reading ⥥ Chameleon is a unique model: it attempts to scale early fusion 🤨 But what is early fusion? Modern vision language models use a vision encoder with a projection layer to project image embeddings so it can be promptable to text decoder (LLM) Early fusion on the other hand attempts to fuse all features together (image patches and text) by using an image tokenizer and all tokens are projected into a shared space, which enables seamless generation 😏 Authors have also introduced different architectural improvements (QK norm and revise placement of layer norms) for scalable and stable training and they were able to increase the token count (5x tokens compared to Llama 3 which is a must with early-fusion IMO) This model is an any-to-any model thanks to early fusion: it can take image and text input and output image and text, but image generation are disabled to prevent malicious use. One can also do text-only prompting, authors noted the model catches up with larger LLMs (like Mixtral 8x7B or larger Llama-2 70B) and also image-pair prompting with larger VLMs like IDEFICS2-80B (see paper for the benchmarks) Thanks for reading!
6 Comments -
Nick Tarazona, MD
👉🏼 Evaluating the efficacy of few-shot learning for GPT-4Vision in neurodegenerative disease histopathology: A comparative analysis with convolutional neural network model 🤓 Daisuke Ono 👇🏻 https://lnkd.in/e2RK6-xr 🔍 Focus on data insights: - GPT-4Vision accurately recognized staining techniques and tissue origin but struggled with specific lesion identification. - The interpretation of images was notably influenced by the provided textual context, sometimes leading to diagnostic inaccuracies. - Few-shot learning significantly improved GPT-4Vision's diagnostic capabilities, enhancing accuracy from 40% in zero-shot learning to 90% with 20-shot learning. 💡 Main outcomes and implications: - GPT-4Vision faces challenges in independently interpreting histopathological images. - Few-shot learning is a promising approach to enhance GPT-4Vision's performance, especially in neuropathology where acquiring extensive labeled datasets is challenging. 📚 Field significance: - Artificial intelligence advancements like GPT-4Vision have expanded the potential for medical image interpretation. - Comparative analysis with traditional convolutional neural network models provides insights into improving diagnostic accuracy in neurodegenerative disease histopathology. 🗄️: #neurodegenerative #histopathology #GPT-4Vision #fewshotlearning
-
Sathya Krishnan Suresh
My latest research work, "Towards smaller, faster decoder-only transformers: Architectural variants and their implications" is now available on arXiv. In this work, we introduce architectural variations to the decoder-only transformer architecture for easier deployment, faster training and faster inferencing with little drop in performance. Do check it out in the link below and feel free to share your thoughts and insights. https://lnkd.in/gF_A4HeU #research #machinelearning #deeplearning
2 Comments -
Nick Tarazona, MD
👉🏼 Using Large Language Models to Detect Depression From User-Generated Diary Text Data as a Novel Approach in Digital Mental Health Screening: Instrument Validation Study 🤓 Daun Shin 👇🏻 https://lnkd.in/ePwXzTdr 🔍 Focus on data insights: - 📊 User-generated diary entries serve as a valuable source for detecting depression, providing a new avenue for mental health screening. - 🤖 LLMs like GPT-3.5 and GPT-4 exhibit high accuracy in identifying depression, particularly when fine-tuned, showcasing the potential of AI in mental health. - 🧠 The study indicates that qualitative data from emotional diary apps could complement traditional quantitative measures. 💡 Main outcomes and implications: - 🔍 The research demonstrated an impressive accuracy rate of 90.2% in detecting depression using fine-tuned LLMs, emphasizing the effectiveness of advanced AI tools in healthcare. - 📈 High specificity
-
Vicki Boykis
Here's my notes and video for the keynote I gave at PyCon Italia ❤️ 🐍 🇮🇹 yesterday, "Closer to the Metal." The post itself is titled "don't worry about LLMs" and that's really what I was trying to get at with this talk. LLMs are a big topic and there is a lot of demand to build them quickly, but the space is very confusing and inflated with hype. How do we remove the hype and use LLMs to build things we need? We focus our attention on one thing at a time and use the practical engineering and machine learning skills we already have.
5 Comments -
Sadaf Walliyani
#Reka Unveils Reka Core: Its Most Capable Multimodal Language Model Reka, an AI startup based in Sunnyvale, California, has unveiled Core, its latest and most advanced multimodal language model developed to date. Their Core model joins their existing Flash (21B parameters) and Edge (7B parameters) models. Reka describes Core as a frontier-class model that delivers industry-leading performance across a wide range of tasks involving text, images, video, and audio. 🌟 According to Reka's technical report, Core was trained entirely from scratch primarily on NVIDIA H100s using Pytorch. Of note, Reka says the model is not finished training and is still improving. However, results from popular benchmarks so far, show that the model is on par with, or outperforms leading models from OpenAI, Anthropic, and Google. The model demonstrates exceptional reasoning skills, including complex language and math abilities, making it ideal for intricate analysis and problem-solving. 🧠📊 #AI #MultimodalModel #LanguageModel #Reka #ArtificialIntelligence #Innovation
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More