Transcendence: Generative Models Can Outperform The Experts That Train Them
Generative models are trained with the simple objective of imitating the conditional probability distribution induced by the data they are trained on. Therefore, when trained on data generated by humans, we may not expect the artificial model to outperform the humans on their original objectives. In this work, we study the phenomenon of transcendence: when a generative model achieves capabilities that surpass the abilities of the experts generating its data. We demonstrate transcendence by training an autoregressive transformer to play chess from game transcripts, and show that the trained model can sometimes achieve better performance than all players in the dataset. We theoretically prove that transcendence can be enabled by low-temperature sampling, and rigorously assess this claim experimentally. Finally, we discuss other sources of transcendence, laying the groundwork for future investigation of this phenomenon in a broader setting.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
A Taxonomy of Transcendence
Although language models are trained to mimic humans, the resulting systems display capabilities beyond the scope of any one person. To understand this phenomenon, we use a controlled setting to identify properties of th…
Skill GeneralizationSelf-transcendence: Is External Feature Guidance Indispensable for Accelerating Diffusion Transformer Training?
Recent works such as REPA have shown that guiding diffusion models with external semantic features (e.g., DINO) can significantly accelerate the training of diffusion transformers (DiTs). However, the use of pretrained e…
Text-to-Image GenerationHumanoid Robots as Agents of Human Consciousness Expansion
The "Loving AI" project involves developing software enabling humanoid robots to interact with people in loving and compassionate ways, and to promote people' self-understanding and self-transcendence. Currently the proj…
STUN: Structured-Then-Unstructured Pruning for Scalable MoE Pruning
Mixture-of-experts (MoEs) have been adopted for reducing inference costs by sparsely activating experts in Large language models (LLMs). Despite this reduction, the massive number of experts in MoEs still makes them expe…
GSM8KMixture-of-ExpertsMultimodal Conditional Image Synthesis with Product-of-Experts GANs
Existing conditional image synthesis frameworks generate images based on user inputs in a single modality, such as text, segmentation, sketch, or style reference. They are often unable to leverage multimodal user inputs …
DiversityImage GenerationImage-to-Image Translation