Blätter-Navigation

Of­f­re 9 sur 177 du 13/09/2019, 13:37

logo

Max-Planck-Insti­tut für Bil­dungs­for­schung - Depart­ment: Humans and Machi­nes

The Max Planck Insti­tute for Human Deve­lop­ment is dedi­ca­ted to the study of human deve­lop­ment and edu­ca­tion. Rese­ar­chers of various disci­pli­nes – inclu­ding psy­cho­logy, edu­ca­tion, socio­logy and medi­cine, as well as history, eco­no­mics, com­pu­ter sci­ence and mathe­ma­tics – work tog­e­ther on inter­di­sci­pli­nary pro­jects at the Ber­lin Insti­tute. The rese­arch ques­ti­ons they examine include how people make effec­tive decisi­ons even under time pres­sure and infor­ma­tion over­load, how the inter­ac­tion bet­ween beha­viour and brain func­tion chan­ges over the life­span, as well as how human emo­ti­ons change in a his­to­ri­cal con­text and how they have affec­ted the course of history its­elf.

Mas­ter Stu­dent Rese­arch Assi­stant in com­pu­ter sci­ence

Work­ing field:

The pro­po­sed pro­ject aims at using state-of-the-art gene­ra­tive models (like GANs and VAEs) to pro­duce new mul­ti­me­dia arte­facts (e.g., pic­tures, music, text) that are indis­tin­guis­ha­ble from human-gene­ra­ted arte­facts. Pre­vious work has shown that deep lear­ning gene­ra­tive models are able to achieve impres­sive results in a range of domains, inclu­ding pic­tures, music, fashion and speech genera­tion. The cur­rent pro­jects wants to repli­cate these results (par­ti­cu­larly in the image domain), and train and test algo­rithms that can later be used in beha­vio­ral expe­ri­men­tal designs.

Require­ments:

The ideal stu­dent rese­ar­cher will be under­ta­king a Mas­ters in Com­pu­ter Sci­ence, Phy­sics, App­lied Maths, or a rela­ted disci­pline.

The fol­lo­wing are requi­red:
  • Know­ledge of machine lear­ning and sta­tis­ti­cal mode­ling, most likely through Mas­ters level clas­ses and pro­jects.
  • Expe­ri­ence with sui­ta­ble pro­gramming packa­ges such as PyTorch, Ten­sor­Flow, and Keras in Python (pre­fer­red) or R.
  • Know­ledge of recent advan­ces in deep lear­ning gene­ra­tive models, like GANs and VAEs.
  • Expe­ri­ence with algo­rith­mic image pro­ces­sing and clas­si­fi­ca­tion

How to ap­ply:

The Max Planck Society stri­ves for gen­der and diver­sity equa­lity. We wel­come app­li­ca­ti­ons from all back­grounds. The Max Planck Society is com­mit­ted to increa­sing the num­ber of indi­vi­du­als with disa­bi­li­ties in its work­force and the­re­fore encou­ra­ges app­li­ca­ti­ons from such qua­li­fied indi­vi­du­als.