Back­ground

The envi­ron­men­tal con­di­tion of the fish farm sites are impor­tant since they are the basic val­ues for the deter­mi­na­tion of the bal­ance between fish farm­ing activ­i­ties and its recip­i­ent envi­ron­ment. Pro­duc­ing fish in a sus­tain­able way is chal­leng­ing, as low effi­cien­cies increase the envi­ron­men­tal bur­den of aqua­cul­ture as well as reduces prof­itabil­i­ty. By improv­ing both effi­cien­cy and yield the sec­tor could grow with­out increas­ing its envi­ron­men­tal impact. Today deci­sions are main­ly based on empiri­cism rather than accu­rate data. Incor­po­ra­tion of SMART tech­nolo­gies for con­trol­ling the basic para­me­ters aims to pro­vide fish farm­ers with clear­er insight in their fish pro­duc­tion for exam­ple by data on oxy­gen and tem­per­a­ture, feed demand and fish behav­iour. Sen­sors can pro­vide infor­ma­tion that facil­i­tates con­tin­u­ous pro­duc­tion and envi­ron­men­tal improve­ment.

SMART fish farm­ing tech­nolo­gies have the poten­tial to improve the effi­cien­cy, reduce ener­gy use and green­house gas­es emis­sion, labour costs, and fish dis­eases, and can stim­u­late the under­stand­ing of the pro­duc­tion process as well as pro­tect­ing the envi­ron­ment. Although the intro­duc­tion of tech­nolo­gies gen­er­al­ly sup­port mod­ern­iza­tion, the devel­op­ment and use of smart tech­nol­o­gy in Indone­sian aqua­cul­ture is still under­de­vel­oped. At present, we do not have a clear view of the tech­nolo­gies imple­ment­ed, the effect on Indone­sian fish farm­ing sys­tems, and the poten­tial of new tech­nolo­gies.


PhD top­ic
The SMART tech­nolo­gies that are being devel­oped for on-farm dis­ease diag­no­sis, and dis­ease and water/milk qual­i­ty mon­i­tor­ing can be sup­port­ed with Big Data ana­lyt­ics for dis­cov­er­ing new knowl­edge. Ani­mal health infor­ma­tion, as well as the diag­no­sis of indi­vid­ual cows with dis­eases, can be deter­mined using the state-of-the-art machine learn­ing meth­ods, tech­niques, and tools. In this research, the data that will be gen­er­at­ed with­in the sys­tem infra­struc­ture will be uti­lized using machine learn­ing approach­es.

First, after data clean­ing, har­mo­niza­tion and inte­gra­tion, and deter­min­ing base­line dis­ease inci­dence and milk qual­i­ty, spa­tial and tem­po­ral devi­a­tions from this base­line (indica­tive for dis­ease out­breaks) are deter­mined. Devi­a­tions are inves­ti­gat­ed to define the thresh­old at which an ear­ly warn­ing should be gen­er­at­ed by tak­ing into account the required lev­els of accu­ra­cy, sen­si­tiv­i­ty, and speci­fici­ty.

Sec­ond, data from milk sam­ples are used to train machine learn­ing algo­rithms for the pre­dic­tion of devi­a­tions in milk com­po­si­tion and mas­ti­tis.

Third, the same is done for water qual­i­ty mon­i­tor­ing, where the nutri­tion­al sta­tus of the pond and activ­i­ty of the fish is mon­i­tored, and ear­ly warn­ing sys­tems are installed that iden­ti­fy under- or over-nutri­fi­ca­tion or deviant fish behav­ioral pat­terns.

Fourth, the robust­ness of dif­fer­ent types of ponds (e.g. shrimp ver­sus fish, nutri­tious ponds ver­sus reg­u­lar ponds, brack­ish ver­sus fresh­wa­ter) against tem­po­ral and spa­tial devi­a­tions is analysed, mak­ing it pos­si­ble to define the best sys­tem under dif­fer­ent cir­cum­stances. Machine learn­ing meth­ods, tech­niques, and tools, derived from the lat­est devel­op­ments in Big Data ana­lyt­ics, will be used in this research. Spe­cif­ic atten­tion will be giv­en to trans­form­ing Big Data mod­els into oper­a­tional deci­sions and deploy­ing mod­els to improve sus­tain­abil­i­ty out­comes (e.g., defin­ing the best feed­ing regime in terms of avail­abil­i­ty of nutri­ents, pond water qual­i­ty and fish pro­duc­tion). Expect­ed out­comes are insight in milk qual­i­ty, water qual­i­ty and epi­demi­ol­o­gy of the dis­ease, eval­u­a­tion of the impact of the inter­ven­tions on farm­ing sys­tem sta­bil­i­ty, and four pub­li­ca­tions for sci­en­tif­ic inter­na­tion­al jour­nals.

Required back­ground

  • A BSc/MSc degree in Com­put­er Sci­ence or any relat­ed dis­ci­pline.
  • Expe­ri­ence in image processing/computer vision tech­niques
  • Expe­ri­ence with machine learn­ing plat­forms such as scik­it-learn, Ten­sor­Flow, and Keras
  • Expe­ri­ence in pro­gram­ming lan­guages such as Python, R, and Java
  • Proven expe­ri­ence with data ana­lyt­ics
  • Excel­lent oral and writ­ing com­mu­ni­ca­tion skills in Eng­lish
  • Pas­sion for doing research with farm­ers and oth­er stake­hold­ers
  • Enthu­si­asm in work­ing in an inter­na­tion­al and inter­dis­ci­pli­nary team and in two coun­tries (Indone­sia and the Nether­lands)


Fur­ther infor­ma­tion
The PhD will be obtained from Wagenin­gen Uni­ver­si­ty, the Nether­lands. The research will be divid­ed between Indone­sia (30 months) and Wagenin­gen Uni­ver­si­ty, the Nether­lands (18 months). Research costs includ­ing visa, trav­el costs for trav­el­ling with­in the project, and costs asso­ci­at­ed with data col­lec­tion are cov­ered by the project. For the peri­od in Wagenin­gen, a grant is avail­able to cov­er salary costs. Salary costs in Indone­sia are not cov­ered in the project, so we have a strong pref­er­ence for can­di­dates that can sup­port them­selves or are sup­port­ed by their employ­er. How­ev­er, in the cir­cum­stance where a can­di­date is unable to find this sup­port but has out­stand­ing cre­den­tials, the project team will con­sid­er sup­port­ing him or her dur­ing their research time in Indone­sia.

In order to be admit­ted to the Wagenin­gen Uni­ver­si­ty PhD pro­gramme, select­ed can­di­dates will have to go through the admis­sion pro­ce­dure. A proven suf­fi­cient lev­el of Eng­lish pro­fi­cien­cy as indi­cat­ed by an IELTS score of 6.5 (with a min­i­mum of 6.0 for speak­ing) is required.

Next to this posi­tion, four oth­er PhD posi­tions (with dif­fer­ent top­ics) are open in the project. Of these five posi­tions, three will be sup­port­ed by an LPDP schol­ar­ship. The LPDP schol­ar­ship is a per­son­al grant that will cov­er all salary costs, both in Indone­sia and in the Nether­lands. Select­ed can­di­dates are encour­aged to enrol in the LPDP schol­ar­ship appli­ca­tion. They will have to go through an addi­tion­al pro­ce­dure to com­pete for the LPDP grant.

For more infor­ma­tion please con­tact prof. dr Ynte Schukken (ynte.schukken@wur.nl) or dr Yeni Herdiyeni
(yeni.herdiyeni@apps.ipb.ac.id)

To show your inter­est, please sub­mit your CV and a let­ter of moti­va­tion before August 23, 2020, to Dr Mar­jolein Derks (marjolein.derks@wur.nl)