Project Florence

Project Florence is a speculative glimpse into our future, where both our natural and digital worlds could co-exist in harmony through enhanced communication. Combining natural language research, biology, design, and engineering, we have created the first instantiation of a plant to human interface through the power of language. In this system, the user first attempts to communicate with or influence the plant through modulated natural language using a Surface computer. The conversation received from the human is analyzed for sentiment and semantic content using a Natural Language Algorithm. The resulting analysis is interpreted into signals that are used to modulate a light source that projects onto the plant. During this conversational turn, the chemical and electrical reactions from the plant and information about its environment are observed. Those signals from the plant are combined with the linguistic input to generate a conversational response from the plant. The resulting response back to the human is a transformation of the original input, driven by linguistic trees as well as lexical paraphrases, with influence based on the data from its sensors. Although the plant does not fully understand the linguistic information, the resulting output is affected in novel and unpredictable ways based on its state. Project Florence provides a base for future research and applications for environmental studies, biological sensors and gen-expressions and new opportunities to bridge our digital and natural worlds together through computation.

Date:
Speakers:
Helene Steiner
Affiliation:
Microsoft Research UK

Series: Microsoft Research Talks