計劃進修PhD/Research討論區 (3)

1001 回覆
7 Like 0 Dislike
2017-03-10 14:21:13
psy/sosci
maths
econ
engine?
仲有無
有無it/computing

Machine Learning/vision算唔算

我個ex而家好似都準備玩machine learning

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field
2017-03-10 14:47:29
psy/sosci
maths
econ
engine?
仲有無
有無it/computing

Machine Learning/vision算唔算

我個ex而家好似都準備玩machine learning

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係
2017-03-10 15:33:41
psy/sosci
maths
econ
engine?
仲有無
有無it/computing

Machine Learning/vision算唔算

我個ex而家好似都準備玩machine learning

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex
2017-03-10 15:37:38

Machine Learning/vision算唔算

我個ex而家好似都準備玩machine learning

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長
2017-03-10 15:40:27
psy/sosci
maths
econ
engine?
仲有無
有無it/computing

Machine Learning/vision算唔算

我個ex而家好似都準備玩machine learning

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field


我Ex係systems engineering嗰邊
佢下年報PhD 應該入硬top 10
2017-03-10 15:41:21

我個ex而家好似都準備玩machine learning

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻
2017-03-10 16:15:30
psy/sosci
maths
econ
engine?
仲有無
有無it/computing

Machine Learning/vision算唔算

Machine Learning 揮手
2017-03-10 16:18:51

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

係,學左成年,學到傻左,先識啲皮毛

btw psychi and text mining 都可以好多野玩,例如trend analysis, diagnosis accuracy, research pattern, symptom associations.....
2017-03-10 17:16:52

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真
2017-03-10 17:21:44

maching learning
text mining揮手

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

係,學左成年,學到傻左,先識啲皮毛

btw psychi and text mining 都可以好多野玩,例如trend analysis, diagnosis accuracy, research pattern, symptom associations.....

我mean係咩data要through text mining?

patient講/寫既野?
2017-03-10 17:37:11

Hmm or Rnn/lstm?

我想試下rnn/reinforcement learning做image classifer

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml
2017-03-10 17:52:29

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

係,學左成年,學到傻左,先識啲皮毛

btw psychi and text mining 都可以好多野玩,例如trend analysis, diagnosis accuracy, research pattern, symptom associations.....

我mean係咩data要through text mining?

patient講/寫既野?

呢part係有小小novelty係度
但同你估既好接近,不過唔是patient
2017-03-10 17:56:27
其實我地是咪真係應該搞個collaboration
有無巴打有興趣
2017-03-10 18:00:51

我果啲好初級的 psychi field

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d
2017-03-10 18:03:27

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

係,學左成年,學到傻左,先識啲皮毛

btw psychi and text mining 都可以好多野玩,例如trend analysis, diagnosis accuracy, research pattern, symptom associations.....

我mean係咩data要through text mining?

patient講/寫既野?

呢part係有小小novelty係度
但同你估既好接近,不過唔是patient

唔怪得google唔到

social network related? 而家好鬼興

不過巴打怕既話可以唔講 純粹個人好奇
2017-03-10 18:05:52

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d

可唔可以少啲數 會考f左
不過psy果啲真係好rely on stat models,好彩models係機計
2017-03-10 18:08:57

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

係,學左成年,學到傻左,先識啲皮毛

btw psychi and text mining 都可以好多野玩,例如trend analysis, diagnosis accuracy, research pattern, symptom associations.....

我mean係咩data要through text mining?

patient講/寫既野?

呢part係有小小novelty係度
但同你估既好接近,不過唔是patient

唔怪得google唔到

social network related? 而家好鬼興

不過巴打怕既話可以唔講 純粹個人好奇


同disease相關,EMR係其中一種,放入去,㩒個制,出啲diagnosis pattern,呢個係其中一種玩左好耐既野
2017-03-10 18:09:41

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d

淨係識artifical neural neteork
2017-03-10 18:11:43
就咁睇上面幾位巴打都係用ML做research, 唔算係做ML research?
2017-03-10 18:11:45

純好奇, 我見到你地話psychi做text mining,但係我唔係好明兩者有咩關係

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d

去conference d 人通常都問
ok 你ml 出到個咁既model
咁significant value 係點
有冇alternative models, 點compare 嫁
我聽完真係暈

我覺得你岩
最難係點係bias variance tradeoff到拎個平衡
我個lecturer 成日講
我地係唔知點解有個咁既model
但我就按某d 條例要個program 整出黎
不過唔係有statistical learning 底既人都係慣用classical stat 既方式諗野
2017-03-10 18:12:27

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d

可唔可以少啲數 會考f左
不過psy果啲真係好rely on stat models,好彩models係機計

唔係stat既話 大多數都有graphical representation
要多d時間睇 不過未至於卡死曬

加油呀 我女朋友bio field成日聽唔明我講math

追dse math既話herman yeung d片好似唔錯

再唔係就khan/other utube

Stat我都幫你唔到 記得呢個post有巴打做bayes果邊應該解答到多d
2017-03-10 18:12:57
就咁睇上面幾位巴打都係用ML做research, 唔算係做ML research?

我冇咁既能力develop algorithm
所以我學都係學點樣apply 係我d data 到
同d engineer 相處過就知係兩個世界
2017-03-10 18:13:11

學緊用ml analyze neuroimaging data
應該同你差唔多
岩岩有個同事出左cerebral cortex

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d

可唔可以少啲數 會考f左
不過psy果啲真係好rely on stat models,好彩models係機計

唔係stat既話 大多數都有graphical representation
要多d時間睇 不過未至於卡死曬

加油呀 我女朋友bio field成日聽唔明我講math

追dse math既話herman yeung d片好似唔錯

再唔係就khan/other utube

Stat我都幫你唔到 記得呢個post有巴打做bayes果邊應該解答到多d

係我
2017-03-10 18:15:11

我有諗個做呢類, eeg/brain-inspired model

不過0 bio/psy background 所以都係做d ML向算

等offer等到頸都長

都好多冇bio/psy background人做
多數都係engineer 或者cs 人做
其實佢地勁好多
你都係當張圖咁分析
我就學到傻

其實都係math trick多

好多時都係將某某方法擺到落另一個problem solve

咁樣相對易做好多

我都想有自己framework 遲下會pub份算係modify左人地個frame既method 不過對mphil個field幫助唔大就真

就係學背後既理論同stat先學到傻, 當然實際programming 時又有另外既問題

不過家下係興咩都放ml 落去
data 唔夠大都ml

我stat就麻麻 programming普普通通

Psy field好似仲係用好stat既model 可能始終你地要搵個reason behind

而家neural network根本就唔係咁需要prior knowledge, trade off就係sample size要大

Model夠簡單 勝左operator夠多
我唔知可唔可以用三個臭皮匠勝過一個豬腳亮黎形容 同埋成件事好blackbox 有d人唔鐘意咁

同埋data set唔夠多都仲有選擇既 係冇咁出名

Eg. Random forest我冇實測過 但係聽講可以prevent overfit 又唔洗話咁多data

Coding都係google架啦

Math我可能因為讀過下basic numerical analysis
而ml其實theory上都係optimization problem 個人覺得幾似下 所以學得輕鬆d

可唔可以少啲數 會考f左
不過psy果啲真係好rely on stat models,好彩models係機計

唔係stat既話 大多數都有graphical representation
要多d時間睇 不過未至於卡死曬

加油呀 我女朋友bio field成日聽唔明我講math

追dse math既話herman yeung d片好似唔錯

再唔係就khan/other utube

Stat我都幫你唔到 記得呢個post有巴打做bayes果邊應該解答到多d

thx 巴打
2017-03-10 18:15:11
就咁睇上面幾位巴打都係用ML做research, 唔算係做ML research?

我冇咁既能力develop algorithm
所以我學都係學點樣apply 係我d data 到
同d engineer 相處過就知係兩個世界

不過psy用到ML text mining 做都勁,應該幾新,可以pub唔少?
吹水台自選台熱 門最 新手機台時事台政事台World體育台娛樂台動漫台Apps台遊戲台影視台講故台健康台感情台家庭台潮流台美容台上班台財經台房屋台飲食台旅遊台學術台校園台汽車台音樂台創意台硬件台電器台攝影台玩具台寵物台軟件台活動台電訊台直播台站務台黑 洞