推荐系统（十六）多任务学习：腾讯PLE模型（Progressive Layered Extraction model）
推荐系统（十六）多任务学习：腾讯PLE模型（Progressive Layered Extraction model）
推荐系统系列博客：
- 推荐系统（一）推荐系统整体概览
- 推荐系统（二）GBDT+LR模型
- 推荐系统（三）Factorization Machines（FM）
- 推荐系统（四）Field-aware Factorization Machines（FFM）
- 推荐系统（五）wide&deep
- 推荐系统（六）Deep & Cross Network（DCN）
- 推荐系统（七）xDeepFM模型
- 推荐系统（八）FNN模型（FM+MLP=FNN）
- 推荐系统（九）PNN模型（Product-based Neural Networks）
- 推荐系统（十）DeepFM模型
- 推荐系统（十一）阿里深度兴趣网络（一）：DIN模型（Deep Interest Network）
- 推荐系统（十二）阿里深度兴趣网络（二）：DIEN模型（Deep Interest Evolution Network）
- 推荐系统（十三）阿里深度兴趣网络（三）：DSIN模型（Deep Session Interest Network）
- 推荐系统（十四）多任务学习：阿里ESMM（完整空间多任务模型）
- 推荐系统（十五）多任务学习：谷歌MMoE（Multi-gate Mixture-of-Experts ）
PLE模型是腾讯发表在RecSys ’20上的文章，这篇paper获得了recsys’20的best paper award，也算为腾讯脱离技术贫民的大业添砖加瓦了。这篇文章号称极大的缓解了多任务学习中存在的两大顽疾：负迁移（negative transfer）现象和跷跷板（seesaw phenomenon），由此带来了相比较其他MTL模型比较大的性能提升。从论文呈现的实验结果也确实是这样的，但从模型结构上来看，更像是大力出奇迹，即性能的提升是由参数量变多而带来的（仅仅是个人看法~）。这篇paper能拿best paper，一方面是实验结果所呈现出来的比较大的性能提升，另一方面是数据分析做的很好，实验做的也很全，因此看起来工作做的很扎实，这是非常值得学习的地方。
本篇博客依然延续之前博客的大纲，会从动机、模型细节结构、代码实现来介绍PLE模型。
一、动机
说到动机，先来说说多任务学习领域中存在的两大问题：
- 负迁移（negative transfer）：MTL提出来的目的是为了不同任务，尤其是数据量较少的任务可以借助transfer learning（通过共享embedding，当然你也可以不仅共享embedding，再往上共享基层全连接网络等等这些很常见的操作）。但经常事与愿违，当两个任务之间的相关性很弱（比如一个任务是判断一张图片是否是狗，另一个任务是判断是否是飞机）或者非常复杂时，往往发生负迁移，即共享了之后效果反而很差，还不如不共享。
- 跷跷板现象：还是当两个task之间相关性很弱或者很复杂时，往往出现的现象是：一个task性能的提升是通过损害另一个task的性能做到的。这种现象存在很久，PLE论文里给它起了个非常贴切的名字『跷跷板』，想象一下你小时候玩跷跷板的情形吧，胖子把瘦子跷起来。
PLE这篇论文，通过大量的实验发现，当前的MTL模型跷跷板现象非常严重（从数据实验中发现问题，进而提出解决办法，会让你的motivation显得非常扎实）。如下图所示，这里的VTR（View-Through Rate）是有效观看率，其定义是用户观看某个视频超过一定时间即认为是一次有效观看，所以是个二分类任务；VCR（View Completion Ratio）是视频观看完成率，是一个回归任务。下面这个图越靠近右上角说明模型在两个task上表现都比较好，左下角则是都很差，因此明显看出，目前MTL领域中主流的模型基本上都存在跷跷板问题（至少在腾讯视频的数据场景下），表现比较好的也就是上一篇博客中介绍的谷歌的MMoE，但还是存在。
二、PLE模型细节
先来从整体上看看这个模型，大家自行对比下MMoE（参见我的博客推荐系统（十五）多任务学习：谷歌MMoE（Multi-gate Mixture-of-Experts ）），可能就能体会到我前面说PLE性能提升更像是复杂参数所带来的这句话了，因为粗略的看，PLE做了deep化（expert完后再来一层expert），显然要比浅层的效果来得好。
我们来自底向上的拆解下PLE：
2.1、Extraction Network
PLE这里相比较MMoE做了比较大的创新，在MMoE里，不同task通过gate（网络）共享相同的expert（网络），而PLE中则把expert分为了两种：共享的expert（即上图中的experts Shared）和每个task单独的expert（task-specific experts）。因此，这种设计即保留了transfer learning（通过共享expert）能力，又能够避免有害参数的干扰（避免negative transfer）。
同样的，在gate网络部分，也分为了task-specific和Shared，以上图左边的这个gate（即上图中红色的gate）为例，它的输入有两部分，分别为Experts A和Experts Shared。而shared gate（上图中蓝颜色的gate）的输入则为三部分：Experts A、Experts Shared和Experts B。最终三部分直接作为下一层的输入，分别对应到下一层的Experts A、Experts Shared和Experts B。有一说一，图figure5如果把中间那个extraction network去掉，会更加清晰。算了，原来想偷个懒，不想画图的，我还是上个简化版的图吧。
再来总结下一些细节，有助于我们代码实现。
- Gate网络的数量取决于task数量，第一层由于多了个shared gate，所以数量等于task数量+1，第二层gate网络数量与task数量相同。Gate网络最后一层全连接层的隐藏单元（即输出）size必须等于expert个数。另外，Gate网络最后的输出会经过softmax进行归一化。
- gate网络作用机制与MMoE相同，以task A为例：输出维度（假设为 ( g 1 , g 2 , g 3 , g 4 , g 5 ) (g_1, g_2, g_3, g_4, g_5) (g1,g2,g3,g4,g5)）等于expert数量(3个task-specific expert，2个shared expert)，意味着 g 1 g_1 g1广播作用于 e x p e r t 1 expert_1 expert1的每一个输出元素上， g 2 g_2 g2广播作用于 e x p e r t 2 expert_2 expert2的每一个输出元素上，以此类推。然后把每一个expert输出向量做element-wise add，即对应位置元素想加，最终得到第二层的expert A。 此外，第二层的Expert shared由第一层全部的expert做element-wise add得到。（建议结合上图理解）
- 每个task的expert数量以及shared expert是个超参，比如最后paddle代码里，每个task有3个expert，shared expert数量为2。实际上每个task的expert数量为3+2=5。
- 相比较MMoE，PLE除了做了一些创新后，网络结构上深度变深了，变成了2层，这也是为什么我说性能提升像是通过增加参数带来的。
- PLE第二层gate网络数量与task数量相同，第一层多了一个shared gate。
2.2、其他一些细节
- 腾讯视频rerank模块的公式为： s c o r e = P V T R W V T R × P V C R W V C R × P S H R W S H R × . . . × P C M R W C M R × f ( v i d e o _ l e n ) score={P_{VTR}}^{W_{VTR}} \times {P_{VCR}}^{W_{VCR}} \times {P_{SHR}}^{W_{SHR}} \times ... \times{P_{CMR}}^{W_{CMR}} \times f(video\_len) score=PVTRWVTR×PVCRWVCR×PSHRWSHR×...×PCMRWCMR×f(video_len)，其中，VCR（View Completion Ratio）为视频观看完成率，VTR（View- Through Rate）为视频是否为有效观看，CMR（Comment Rate）为评论率，SHR（Share Rate）为分享率。 W W W为权重因子，用于调整每个指标的权重。
- 关于loss函数，每个task的loss加了个可学习的权重参数 w w w，用于模型自动学习每个task的loss的权重。
三、代码实现
老规矩，直接上paddle给出的实现代码，我这里把向量维度加了注释，方便理解。
组网代码：
import paddle
import paddle.nn as nn
import paddle.nn.functional as F
class PLELayer(nn.Layer):
def __init__(self, feature_size, task_num, exp_per_task, shared_num,
expert_size, tower_size, level_number):
super(PLELayer, self).__init__()
"""
feature_size: 499
"""
self.task_num = task_num # 2
self.exp_per_task = exp_per_task # 3
self.shared_num = shared_num # 2
self.expert_size = expert_size # 16
self.tower_size = tower_size # 8
self.level_number = level_number # 2
# ple layer
"""
ple_layers:
[SinglePLELayer(
(lev_0_exp_0_0): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_0_0)
(lev_0_exp_0_1): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_0_1)
(lev_0_exp_0_2): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_0_2)
(lev_0_exp_1_0): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_1_0)
(lev_0_exp_1_1): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_1_1)
(lev_0_exp_1_2): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_1_2)
(lev_0_exp_shared_0): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_shared_0)
(lev_0_exp_shared_1): Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_shared_1)
(lev_0_gate_0): Linear(in_features=499, out_features=5, dtype=float32, name=lev_0_gate_0)
(lev_0_gate_1): Linear(in_features=499, out_features=5, dtype=float32, name=lev_0_gate_1)
(lev_0_gate_shared_): Linear(in_features=499, out_features=8, dtype=float32, name=lev_0_gate_shared_)
(_param_gate_shared): Linear(in_features=499, out_features=8, dtype=float32, name=lev_0_gate_shared_)),
SinglePLELayer(
(lev_1_exp_0_0): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_0_0)
(lev_1_exp_0_1): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_0_1)
(lev_1_exp_0_2): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_0_2)
(lev_1_exp_1_0): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_1_0)
(lev_1_exp_1_1): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_1_1)
(lev_1_exp_1_2): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_1_2)
(lev_1_exp_shared_0): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_shared_0)
(lev_1_exp_shared_1): Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_shared_1)
(lev_1_gate_0): Linear(in_features=16, out_features=5, dtype=float32, name=lev_1_gate_0)
(lev_1_gate_1): Linear(in_features=16, out_features=5, dtype=float32, name=lev_1_gate_1))]
"""
self.ple_layers = []
for i in range(0, self.level_number):
if i == self.level_number - 1:
ple_layer = self.add_sublayer(
name='lev_' + str(i),
sublayer=SinglePLELayer(
feature_size, task_num, exp_per_task, shared_num,
expert_size, 'lev_' + str(i), True))
self.ple_layers.append(ple_layer)
break
else:
ple_layer = self.add_sublayer(
name='lev_' + str(i),
sublayer=SinglePLELayer(
feature_size, task_num, exp_per_task, shared_num,
expert_size, 'lev_' + str(i), False))
self.ple_layers.append(ple_layer)
feature_size = expert_size
# task tower
self._param_tower = []
self._param_tower_out = []
for i in range(0, self.task_num):
# [16, 8]
linear = self.add_sublayer(
name='tower_' + str(i),
sublayer=nn.Linear(
expert_size,
tower_size,
weight_attr=nn.initializer.Constant(value=0.1),
bias_attr=nn.initializer.Constant(value=0.1),
#bias_attr=paddle.ParamAttr(learning_rate=1.0),
name='tower_' + str(i)))
self._param_tower.append(linear)
# [8, 2]
linear = self.add_sublayer(
name='tower_out_' + str(i),
sublayer=nn.Linear(
tower_size,
2,
weight_attr=nn.initializer.Constant(value=0.1),
bias_attr=nn.initializer.Constant(value=0.1),
name='tower_out_' + str(i)))
self._param_tower_out.append(linear)
def forward(self, input_data):
"""
input_data: Tensor(shape=[2, 499], 2--->batch_size
"""
inputs_ple = []
# task_num part + shared part
# task_num-->2
for i in range(0, self.task_num + 1):
inputs_ple.append(input_data)
# multiple ple layer
ple_out = []
# level_number-->2
for i in range(0, self.level_number):
print "^^^^^number: %d^^^^^^"%i
ple_out = self.ple_layers[i](inputs_ple)
inputs_ple = ple_out
# ple_out, [Tensor(shape=[2, 16], Tensor(shape=[2, 16]]
#assert len(ple_out) == self.task_num
output_layers = []
for i in range(0, self.task_num):
# Tensor(shape=[2, 8]
cur_tower = self._param_tower[i](ple_out[i])
cur_tower = F.relu(cur_tower)
out = self._param_tower_out[i](cur_tower)
out = F.softmax(out)
out = paddle.clip(out, min=1e-15, max=1.0 - 1e-15)
output_layers.append(out)
return output_layers
class SinglePLELayer(nn.Layer):
def __init__(self, input_feature_size, task_num, exp_per_task, shared_num,
expert_size, level_name, if_last):
super(SinglePLELayer, self).__init__()
# input_feature_size: first layer: 499, second layer=expert_size-->16
self.task_num = task_num # 2
self.exp_per_task = exp_per_task # 3
self.shared_num = shared_num # 2
self.expert_size = expert_size # 16
self.if_last = if_last # 1
self._param_expert = []
# task-specific expert part
for i in range(0, self.task_num):
for j in range(0, self.exp_per_task):
# [499, 16]
linear = self.add_sublayer(
name=level_name + "_exp_" + str(i) + "_" + str(j),
sublayer=nn.Linear(
input_feature_size,
expert_size,
weight_attr=nn.initializer.Constant(value=0.1),
bias_attr=nn.initializer.Constant(value=0.1),
name=level_name + "_exp_" + str(i) + "_" + str(j)))
self._param_expert.append(linear)
# shared expert part
for i in range(0, self.shared_num):
# [499, 16]
linear = self.add_sublayer(
name=level_name + "_exp_shared_" + str(i),
sublayer=nn.Linear(
input_feature_size,
expert_size,
weight_attr=nn.initializer.Constant(value=0.1),
bias_attr=nn.initializer.Constant(value=0.1),
name=level_name + "_exp_shared_" + str(i)))
self._param_expert.append(linear)
# task gate part
self._param_gate = []
cur_expert_num = self.exp_per_task + self.shared_num
for i in range(0, self.task_num):
# [499, 5]
linear = self.add_sublayer(
name=level_name + "_gate_" + str(i),
sublayer=nn.Linear(
input_feature_size,
cur_expert_num,
weight_attr=nn.initializer.Constant(value=0.1),
bias_attr=nn.initializer.Constant(value=0.1),
name=level_name + "_gate_" + str(i)))
self._param_gate.append(linear)
# shared gate
if not if_last:
# 8
cur_expert_num = self.task_num * self.exp_per_task + self.shared_num
# [499, 8]
linear = self.add_sublayer(
name=level_name + "_gate_shared_",
sublayer=nn.Linear(
input_feature_size,
cur_expert_num,
weight_attr=nn.initializer.Constant(value=0.1),
bias_attr=nn.initializer.Constant(value=0.1),
name=level_name + "_gate_shared_"))
self._param_gate_shared = linear
def forward(self, input_data):
print "****SinglePLELayer forward****"
expert_outputs = []
# task-specific expert part
# task_num: 2
# exp_per_task: 3
"""
_param_expert: level0:
[Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_0_0),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_0_1),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_0_2),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_1_0),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_1_1),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_1_2),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_shared_0),
Linear(in_features=499, out_features=16, dtype=float32, name=lev_0_exp_shared_1)]
level1:
[Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_0_0),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_0_1),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_0_2),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_1_0),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_1_1),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_1_2),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_shared_0),
Linear(in_features=16, out_features=16, dtype=float32, name=lev_1_exp_shared_1)]
"""
# [0, 1]
for i in range(0, self.task_num):
for j in range(0, self.exp_per_task): # [0,1,2]
# 0,1,2, 2,3,4
linear_out = self._param_expert[i * self.task_num + j]( # bug, task_num-->exp_per_task
input_data[i])
# Tensor(shape=[2, 16]
expert_output = F.relu(linear_out)
expert_outputs.append(expert_output)
# shared expert part
# shared_num: 2
# [0, 1]
for i in range(0, self.shared_num):
# self.exp_per_task * self.task_num = 3*2
linear_out = self._param_expert[self.exp_per_task * self.task_num + i](input_data[-1])
# Tensor(shape=[2, 16]
expert_output = F.relu(linear_out)
expert_outputs.append(expert_output)
# task gate part
outputs = []
"""
self._param_gate:
[Linear(in_features=499, out_features=5, dtype=float32, name=lev_0_gate_0),
Linear(in_features=499, out_features=5, dtype=float32, name=lev_0_gate_1)]
"""
# [0, 1]
for i in range(0, self.task_num):
# 5
cur_expert_num = self.exp_per_task + self.shared_num
# Tensor(shape=[2, 5]
linear_out = self._param_gate[i](input_data[i])
cur_gate = F.softmax(linear_out)
# Tensor(shape=[2, 5, 1]
cur_gate = paddle.reshape(cur_gate, [-1, cur_expert_num, 1])
# f^{k}(x) = sum_{i=1}^{n}(g^{k}(x)_{i} * f_{i}(x))
# expert_outputs[0:3], expert_outputs[3:6]
# expert_outputs[-2:]
cur_experts = expert_outputs[i * self.exp_per_task:(i + 1) * self.exp_per_task] + \
expert_outputs[-int(self.shared_num):]
# Tensor(shape=[2, 80])
expert_concat = paddle.concat(x=cur_experts, axis=1)
# [2, 5, 16]
expert_concat = paddle.reshape(
expert_concat, [-1, cur_expert_num, self.expert_size])
# Tensor(shape=[2, 5, 16]
# [2, 5, 16] * Tensor(shape=[2, 5, 1])
cur_gate_expert = paddle.multiply(x=expert_concat, y=cur_gate)
# Tensor(shape=[2, 16]
cur_gate_expert = paddle.sum(x=cur_gate_expert, axis=1)
outputs.append(cur_gate_expert)
# shared gate
if not self.if_last:
# 8
cur_expert_num = self.task_num * self.exp_per_task + self.shared_num
# Tensor(shape=[2, 8], [2, 499] * [499, 8]
linear_out = self._param_gate_shared(input_data[-1])
cur_gate = F.softmax(linear_out)
# Tensor(shape=[2, 8, 1]
cur_gate = paddle.reshape(cur_gate, [-1, cur_expert_num, 1])
cur_experts = expert_outputs
# Tensor(shape=[2, 128]
expert_concat = paddle.concat(x=cur_experts, axis=1)
# Tensor(shape=[2, 8, 16])
expert_concat = paddle.reshape(
expert_concat, [-1, cur_expert_num, self.expert_size])
# Tensor(shape=[2, 8, 16]), --->[2, 8, 16]*[2, 8, 1]
cur_gate_expert = paddle.multiply(x=expert_concat, y=cur_gate)
# Tensor(shape=[2, 16])
cur_gate_expert = paddle.sum(x=cur_gate_expert, axis=1)
outputs.append(cur_gate_expert)
return outputs
loss函数代码：
def net(self, inputs, is_infer=False):
input_data = inputs[0]
label_income = inputs[1]
label_marital = inputs[2]
PLE = PLELayer(self.feature_size, self.task_num, self.exp_per_task,
self.shared_num, self.expert_size, self.tower_size,
self.level_number)
pred_income, pred_marital = PLE.forward(input_data)
pred_income_1 = paddle.slice(
pred_income, axes=[1], starts=[1], ends=[2])
pred_marital_1 = paddle.slice(
pred_marital, axes=[1], starts=[1], ends=[2])
auc_income, batch_auc_1, auc_states_1 = paddle.static.auc(
#auc_income = AUC(
input=pred_income,
label=paddle.cast(
x=label_income, dtype='int64'))
#auc_marital = AUC(
auc_marital, batch_auc_2, auc_states_2 = paddle.static.auc(
input=pred_marital,
label=paddle.cast(
x=label_marital, dtype='int64'))
if is_infer:
fetch_dict = {'auc_income': auc_income, 'auc_marital': auc_marital}
return fetch_dict
cost_income = paddle.nn.functional.log_loss(
input=pred_income_1,
label=paddle.cast(
label_income, dtype="float32"))
cost_marital = paddle.nn.functional.log_loss(
input=pred_marital_1,
label=paddle.cast(
label_marital, dtype="float32"))
avg_cost_income = paddle.mean(x=cost_income)
avg_cost_marital = paddle.mean(x=cost_marital)
cost = avg_cost_income + avg_cost_marital
self._cost = cost
fetch_dict = {
'cost': cost,
'auc_income': auc_income,
'auc_marital': auc_marital
}
return fetch_dict
参考文献
[1]: Tang H , Liu J , Zhao M , et al. Progressive Layered Extraction (PLE): A Novel Multi-Task Learning (MTL) Model for Personalized Recommendations[C]// RecSys '20: Fourteenth ACM Conference on Recommender Systems. ACM, 2020.