summer
deer

An open-source platform for AI creators, scientific workflows, and creative intelligence.

Introduction

What Is SummerDeer

SummerDeer is a commitment of ideas for the incoming summer, a creative informatics platform. Artificial intelligence (AI) has produced amazing progress, from early convolutional neural networks (CNNs) [1] and AlexNet [2] to Transformers [3] and modern large language models (LLMs) [4]. But current systems still face open challenges in robust real-world visual perception [5], sim-to-real and long-horizon robotic control [6], repository-scale programming [7], and reliable planning or decision-making [8]. I believe AI should become more than a tool for generating output. It should become a system that helps us study information, build practical machine learning methods, and create more intelligent workflows. From neuroscience, I believe intelligence is not only about growth. It is also about shaping, pruning, and distilling what matters. Our minds are built not only by what we gain, but also by what we remove.

This Platform is open to contributors, researchers, developers, and young creative minds. Contributors can upload workflows, improve the platform, and receive percentage-based contribution credits within the SummerDeer community. These credits are designed as a knowledge-exchange system: useful workflows create value, and contributors receive recognition for the value they add. Our goal is to build a large workflow network that is more precise, structured, and controllable than ordinary large language model outputs. SummerDeer is not only a website. It is an open intelligence architecture.

Historians write history. Attorneys and entrepreneurs build modern society. And we are scientists — we name the future.
References
  1. Fukushima, K. (1980). Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biological Cybernetics.
  2. Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet Classification with Deep Convolutional Neural Networks. NeurIPS.
  3. Vaswani, A., Shazeer, N., Parmar, N., et al. (2017). Attention Is All You Need. NeurIPS.
  4. Brown, T. B., Mann, B., Ryder, N., et al. (2020). Language Models are Few-Shot Learners. NeurIPS.
  5. Hendrycks, D., & Dietterich, T. (2019). Benchmarking Neural Network Robustness to Common Corruptions and Perturbations. ICLR.
  6. Salvato, E., Fenu, G., Medvet, E., & Pellegrino, F. A. (2021). Crossing the Reality Gap: A Survey on Sim-to-Real Transferability of Robot Controllers in Reinforcement Learning. IEEE Access.
  7. Jimenez, C. E., Yang, J., Wettig, A., Yao, S., Pei, K., Press, O., & Narasimhan, K. (2024). SWE-bench: Can Language Models Resolve Real-World GitHub Issues? ICLR.
  8. Valmeekam, K., Marquez, M., Olmo, A., Sreedharan, S., & Kambhampati, S. (2023). PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Change. NeurIPS.

Why SummerDeer

Large Repository Models are less popular than LLMs, but they may be better at capturing human preferences. Modern AI tools can generate code, but they often do not understand why a person prefers one process over another. Workflows provide a clearer way to learn preference through input-output patterns, instead of relying only on question-answer embeddings.

Nordic Concision is the key value of SummerDeer. Inspired by the Law of Jante, this platform values clear thinking and shared contribution. Beauty is not born from complexity, but from simplicity.

Ideological thinking. We need new AI ideas, and we need the courage to make them real. The value of a website is not only in sharing methodologies, but in building an interactive space where brilliant minds from different fields can work together. I am the owner only for now. After SummerDeer is published, everyone who contributes to the platform will become part of its ownership. Voting will be the core way we make decisions and shape SummerDeer into a better place.

Contents

Mathematical Design

  • Workflow-based reasoning structures
  • Input-output embedding maps
  • Optimization and decision loops
  • Practical models for scientific prediction

Computer Vision

  • Image understanding and visual tracking
  • Object detection and spatial recognition
  • Real-world visual perception systems
  • Vision tools for scientific data analysis

Document Creator

  • Scientific notes and structured writing
  • Automatic summaries from files and papers
  • Research reports, figures, and web content
  • Knowledge archives for long-term memory

Links

Unimap

A browser-based embedding map for exploring files, chunks, similarity, clusters, and workflow knowledge structure.

Platform

Workflow Intelligence

SummerDeer connects practical workflows, file reading, embedding maps, and scientific tool design into one open platform. The goal is to build a workflow network that is clearer, more controllable, and more useful than ordinary text-only AI output.

Core idea: people contribute workflows, workflows become structured knowledge, and structured knowledge becomes the foundation for future intelligent systems.

Uninote

A browser-based universal file reader for notes, documents, code, images, PDFs, and structured files.