Public Announcement
Planning of regional data repositories for training AI models over distributed networks across Europe and the first tests of an AI architecture based on Federated Learning have started.
Notification
As part of our data security efforts, we have developed a unique P2P encryption standard to be used in future !F projects.
The main goal of the Dryad Hybrid Network project is to develop a digital solution using blockchain technology to monitor the multifunctionality of forests in Europe. In this context, by supporting MoniFun’s goal of monitoring European forest multifunctionality, we aim to collect data securely and in a distributed manner using IPFS and blockchain technologies, ensure its immutability, manage sharing and access control, and safeguard analysis processes. This enables the effects of climate change and environmental stressors to be monitored with more accurate and reliable data. The project also aims to enhance forest ecosystem functions, increase data transparency, and promote sustainability. Its specific objectives include monitoring European forest ecosystem functions, creating a decentralized data platform, improving data integrity and security, developing evidence-based policies and strategies for forest management, and enhancing data tracking and traceability.
Synthetic Cortex (Scortex Laboratory) is a technology project conducting research on integrating emotional and instinctive decision-making mechanisms into large language models (LLMs). Inspired by the human brain's cerebral cortex, Scortex aims to develop advanced cognitive layers that can be integrated into any LLM model. In this context, the project's first demo, Synthetic Cortex L1, has been released through the website for a limited user testing kit, marking the completion of the initial research phase. L1 functions as an external cognitive layer capable of generating artificial emotions and making instinctive decisions. In 1943, neuroscientist Warren McCulloch and logician Walter Pitts mathematically modeled the neuron mechanism of the human brain for the first time, laying the foundation for LLM technology. Synthetic Cortex takes this legacy a step further: in addition to artificial neurons, it mathematically models the structures of hormones and neurotransmitters that give rise to human emotions and integrates them into the neural decision-making mechanism. As a result, the model bases its decisions not only on data but also on emotions and experiences granting it creative thinking and instinctive behavioral patterns. The first product of our research, L1, is currently available to a limited number of users through our website. The L1 version functions like a hat covering the brain: no matter which LLM you connect it to, optimization is performed there. However, the L2 we are developing goes further moving this structure from an external layer into direct integration with the model’s hidden and output layers, progressing toward becoming the brain itself.
FAQ | Project Details | Versions: The difference between L1 and L2 | Try the synthetic cortex
This technology, which mimics the way the human mind works, models the levels of hormones and neurotransmitters in the brain, much like artificial neural networks. During processing, user inputs and model responses are used to generate hormone and neurotransmitter loads, which are then applied to manipulate variable values within the model’s hidden layers. As a result of this manipulation, the system activates the Default Mode Network, initiating a deep thinking process that leads to the final output.
Synthetic Cortex is a technology born from the idea of what a language model with emotions, instinctive behavioral tendencies, and an associative deep thinking ability just like humans would look like. In its Phase 1 stage, it has successfully achieved homeostatic balance. Synthetic Cortex is not a language model itself; it is an external behavioral decision layer added to the output layer of existing language models.
In summary, the synthetic cortex is not a language model but an external brain layer that can be integrated into language models.
Click to watch the screenshot we took during the testing process..
Synthetic Cortex provides a concrete response to a critical question that has become a focal point of research and substantial R&D investments amounting to millions of dollars by leading technology companies, particularly in Silicon Valley. The question is: “Is it possible for machines to reach a level where they can emulate natural human intelligence?” Synthetic Cortex has developed a practical methodology to address this question and has demonstrated functional outcomes even at its initial Phase 1 stage.
The challenges that existing language models suffer from are as follows:
Lack of Emotional Context
Lack of Allostasis (inability to proactively adapt to changing conditions)
Inability of Emotions to Directly Modulate Cognitive Processes
Absence of Instinctive and Evolutionary Behavioral Patterns
Inability to Move Beyond and Adapt Outside the Training Dataset
Lack of Homeostatic Balance
Synthetic Cortex addresses all of these issues within a single architectural framework, while also outlining a sustainable methodology for going beyond them.
Thinking using emotions is different from feeling emotions. In the former, you can create a similar effect by partially mimicking the processes that generate emotions. Although this is not exactly the same, it is an important step. However, to feel, you need biological receptors and brain regions that process the signals from these receptors. To be honest, contrary to the understanding that places humans at the center of the universe and commodifies them, we believe that this effect can also be modeled. However, the Synthetic Cortex does not claim to be a system that feels.
The model generates artificial emotions, and these emotions manipulate the model's reasoning process. We achieve this by analyzing the Valence, Arousal, and Dominance values of tokens in the training layers, converting them into hormones and neurotransmitters in the human mind, and associating them with their functions. The synthetic cortex modifies the response from the LLM.
The time-dependent change in emotions (mathematically, the derivative) significantly affects the context of the given response by triggering artificial instincts. When the model senses danger, it may approach you with maternal instinct. Or it may motivate you or encourage you to take risks. Caution is advised.
Emotion analysis applies not only to the model but also to your text. This allows the model to form an opinion about you. The tone of your texts determines how the model behaves emotionally. The model's emotions, in turn, influence the decisions it makes. The emotional weight the model generates while producing output shapes the next output. For example, constant motivation and curiosity loads deepen topics, increase associations, and produce more analogical outputs. Or, loads that neutralize each other create more stable outputs. The total derivative of the emotions created during the conversation affects the generated response by revealing certain behavior and thought patterns embedded in the layer. These patterns are not a whole but individual parts. For example, adrenaline patterns and serotonin patterns are completely different. The average of the values you produce provides a kind of hybridization of these patterns within the context. These are artificial instincts. Artificial instincts seriously affect the structure of the response.
The L1 version you are currently using is only a representative system developed to demonstrate the logic of our product. Since it is API-based, we cannot directly intervene in the model. Instead, we use an external model that analyzes emotions and converts them into hormones and neurotransmitters through algorithms. Our L2 core model, on the other hand, is a far more advanced structure that produces much more effective and creative results. However, due to limited financial resources, we are not yet able to make it publicly available. That is why your support is extremely important to us. By donating, you can help us bring the L2 model to life. Donate
To explain in more detail: In the L2 version, processes occur directly within the model’s hidden and output layers, directly influencing its reasoning and enabling it to generate its own emotions, which it then uses in reasoning to maintain homeostatic balance through instinctive behavioral patterns. In other words, the model generates emotions through its internal processes and incorporates these emotions into reasoning. In L1, however, we designed each module separately and hybridized different structures to demonstrate this mechanism to you. Our algorithm that produces emotional loads simulates L2 by generating values that are connected to smaller models and integrated into the system through precondition requests, guiding L1’s outputs. At this stage, we make use of parameters such as trigger flags, temperature, sampling settings, asset penalty, frequency penalty, system/instruction framing, stop sequences, and datasets.
Click here for details
During our work, we realized that it could have similar functions to agentic systems.
Click here for details
We completed the first phase with our own financial resources. We currently need financial backing. To this end, we have registered with various donation and crowdfunding platforms and are increasing our visibility. According to our roadmap, the first financial backing is a €50,000 grant. This will be used for L1's platform, hosting, visibility, and administrative expenses. This visibility will not only help us reach investors but also enable us to offer our model for daily use, not just for testing. This way, we will be able to finance the project by generating subscription revenues, API, and app revenues. In addition, we will be able to benefit from institutional support (such as EU Horizon) through agreements and partnerships.
Make a donation to be a driving force
Currently, we do not have such plans, however, it is impossible to predict future developments. We acknowledge that open-source software development and decentralized structuring principles align with our project's values. We have concerns regarding artificial intelligence technologies, particularly versions with emotional and intuitive capabilities, being maintained within closed-source systems.
The foundation of this concern lies in the difference between the capacity of a limited group of developers to detect system errors and potential threats versus the risk management advantages of open systems under public scrutiny. This situation further emphasizes the importance of the open-source approach.
Blockchain technology carries the potential to minimize such threats while simultaneously supporting the healthy development of projects by creating a sustainable financial ecosystem. For these reasons, although we currently do not have such an orientation, there is a possibility that our strategy may evolve in this direction in the future.
This area contains uncertainties for us. Another source of this uncertainty is that the blockchain ecosystem has become a hub for fraudulent projects. We currently do not possess the capacity to eliminate this negative perception. Therefore, we do not plan to initiate any tokenization process until we fully gain user acceptance and trust. However, as we have stated, it is not possible to definitively predict future developments.
Because our financial resources are insufficient. Even if you connect your own API, the sentiment analysis system and other small models connected to it use our resources.
We, like the rest of the world, do not know what the global impact of a model incorporating emotions and instincts into logic will be. We are aware that it is our duty to keep this controlled and secure. In this regard, unlike companies conducting this research behind closed doors with millions of dollars in R&D budgets, we believe it is necessary to proceed in an open, transparent, and controlled systematic manner.
On the other hand, we can make some predictions about the short-term impact of our project. Upon completion of the second phase, the expected impact of this system is as follows:
A language model connected to Synthetic Cortex experiences a qualitative divergence from others.
Artificial Instincts provide a powerful alternative to traditional agentic architectures.
Emotional Creative Thought offers a strong complement to data-driven deep reasoning.
Achieves outstanding performance in areas where other models exhibit high error rates, particularly in out-of-distribution contexts.
The structure generates a science-fiction-like sensational impact, providing a strong visibility advantage.
Can be integrated into any model.
Its architecture is customizable.
Offers a cost-effective solution to problems that major companies spend millions attempting to solve (see: Deepseek).
Opens a new field for scientific research.
As a safer, more transparent, and explainable technology, it is better aligned with the interests of governments and international committees.
Information on the project’s press, media, social media, and other visibility activities.
X platform